Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of riparian buffer width on songbirds and forest structure in the southern interior of British.. Haag, Devon Adaire 2002-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
ubc_2002-0414.pdf [ 5.34MB ]
Metadata
JSON: 1.0090468.json
JSON-LD: 1.0090468+ld.json
RDF/XML (Pretty): 1.0090468.xml
RDF/JSON: 1.0090468+rdf.json
Turtle: 1.0090468+rdf-turtle.txt
N-Triples: 1.0090468+rdf-ntriples.txt
Original Record: 1.0090468 +original-record.json
Full Text
1.0090468.txt
Citation
1.0090468.ris

Full Text

E F F E C T S OF R I P A R I A N B U F F E R W I D T H O N SONGBIRDS A N D F O R E S T S T R U C T U R E IN T H E S O U T H E R N I N T E R I O R O F BRITISH C O L U M B I A by  D E V O N A D AIRE H A A G B.Sc, University of British Columbia, 1994  A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T OF T H E REQUIREMENTS F O R T H E D E G R E E OF M A S T E R OF S C I E N C E In  T H E F A C U L T Y OF G R A D U A T E S T U D I E S Department of Forest Sciences We accept this thesis as conforming to the required standard  T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A October 2002 © Devon Adaire Haag, 2002  In  presenting  degree freely  at  this  the  available  copying  of  department publication  of  in  partial  fulfilment  of  the  University  of  British  Columbia,  I  agree  for  this or  thesis  reference  thesis by  this  for  his thesis  and  study.  scholarly  or  her  for  financial  Department  Date  DE-6  (2788)  Columbia  00,3  purposes  gain  shall  that  agree  may  representatives.  permission.  T h e U n i v e r s i t y o f British Vancouver, Canada  I further  requirements  It not  be is  that  the  Library  permission  granted  by  understood be  for  allowed  an  advanced  shall for  the that  without  make  it  extensive  head  of  my  copying  or  my  written  ABSTRACT  In harvested landscapes, the retention of riparian buffers along streams may mitigate the effects of habitat loss and fragmentation by providing usable habitat for songbirds. To explore this hypothesis, I studied the influence of riparian buffer width on breeding songbirds and forest structure in a high elevation forest of south-central British Columbia. I studied four different buffer widths, consisting of very narrow (2-3 m), narrow (11-15 m), medium (30-34 m) and wide buffers (57-69 m). Buffer and control (unharvested forest) sites were each replicated twice for a total of ten study sites. I conducted spot map surveys and habitat sampling to measure width effects on songbird density and vegetation, and to assess the influence of forest structure on songbird density. To examine habitat use by forest birds, I observed the foraging behaviours and movements of four songbird species: winter wren, yellow-rumped warbler, golden-crowned kinglet and Townsend's warbler. Riparian buffer width had several effects on the songbirds breeding within the study area. First, buffer width influenced songbird community structure and composition. The juxtaposition of clearcut and forest in the study grids containing medium and wide buffers maximized species richness and diversity. As buffer width decreased, generalist and open-habitat species replaced forest species within the study grids; very narrow and narrow buffers provided little habitat for forest songbirds. Second, although changes in forest structure occurred across buffers, width was the most important factor determining the richness and density of forest songbird species. Third, buffer width influenced the movement patterns of foraging songbirds. Individuals in buffers moved greater distances upstream and downstream than they did towards and away from the stream; individuals in unharvested stands moved almost equally in all directions. Overall, there did not appear to be a threshold buffer width beyond which there was a disproportionate loss of species and individuals. Although several common forest species were present to a certain extent in all riparian buffers, forest songbirds would benefit most from buffers > 30 m in width.  TABLE OF CONTENTS ABSTRACT  ii  LIST OF T A B L E S  v  LIST OF FIGURES  vi  ACKNOWLEDGEMENTS  viii  INTRODUCTION  1  C H A P T E R 1. The effect of riparian buffer width on breeding songbirds  4  INTRODUCTION  4  METHODS Study Area and Experimental Design Breeding Bird Sampling Territory Delineation with GIS Data Analysis RESULTS Community Characteristics Songbird Density  8 8 10 11 12 :  16 16 17  .„  DISCUSSION Width Effects on Community Characteristics Width Effects on Songbird Density Edge Effects Conclusions  ;  20 20 21 26 28  C H A P T E R 2. Vegetation characteristics of riparian buffers and the use of riparian buffers by foraging forest songbirds 42 INTRODUCTION  42  METHODS : Habitat Sampling Behaviour Sampling Data Analysis  45 45 47 48  RESULTS Forest Structure Songbird-Habitat Associations  ....52 52 54  :  DISCUSSION.. ! : Width Effects on Forest Structure Songbirds and Buffer Width vs. Forest Structure Habitat Use and Movements Conclusions...... CONCLUSIONS A N D RECOMMENDATIONS Study Limitations  59 59 60 62 66 : <  82 83  Management Recommendations L I T E R A T U R E CITED  86 89  APPENDIX 1: An example of territory delineation for the Townsend's warbler.... 100 APPENDIX 2: Songbird species list and species abbreviations for the study area.. 101 APPENDIX 3: An example of plot layout for vegetation sampling within a study grid. 104  iv  LIST O F T A B L E S Table 1.1. Ecological characteristics of each study grid  31  Table 1.2. Habitat associations and mean territories/ha (± 1 SE) of songbird species breeding within the study grids in 1999  32  Table 1.3. Similarity values for the Bray-Curtis Community Similarity Index  33  Table 1.4. Results for Kruskal-Wallis tests of treatment effect on non-normally distributed songbird density data (degrees of freedom for each test = 4)  34  Table 2.1. Mean percent cover of ground cover types across treatments and in the unharvested forest  67  Table 2.2. Results of factor analysis with Equamax rotation. Only vegetation variables with component loadings >0.55 are shown  68  Table 2.3. Partial correlation values for the richness and density of forest songbird species vs. each of the four forest structure factors. Buffer width is the variable controlled for in each correlation  69  LIST OF F I G U R E S Figure 1.1. Location of Tree Farm License (TFL) #35 within the Kamloops Forest Region (KFR)  35  Figure 1.2. Location of study sites within T F L #35  36  Figure 1.3. Graphical representation of treatments, depicting the 150 m x 300 m spot-map grids and each buffer width on one side of the stream and continuous forest on the other ; 37 Figure 1.4. Mean species richness (+ 1 SE), and mean Shannon-Wiener and Simpson diversity (+ 1 SE) in study grids across buffer types  38  Figure 1.5. Relationship between riparian buffer width and species richness in the buffers only  39  Figure 1.6. Relationship between riparian buffer width and the territory density (square-root transformed) of (a) all species combined, and (b) forest species, within study grids  40  Figure 1.7. Relationship between riparian buffer width and the territory density (square-root transformed) of (a) golden-crowned kinglets (b) Townsend's warblers (c) winter wrens, and (d) yellow-rumped warblers, within study grids  41  Figure 2.1. (a) Direction angles (relative to stream) used to record songbird movements, and (b) movement classes used for analysis  70  Figure 2.2. (a) Mean basal area (± 1 SE), and (b) mean percent canopy cover (± 1 SE) in forest plots across treatments  71  Figure 2.3. Changes in percent tree species composition with distance from stream in (a) medium buffers, (b) wide buffers, and (c) the unharvested forest  72  Figure 2.4. Mean percent cover of coniferous and deciduous shrubs (± 1 SE) in forest plots across treatments  73  Figure 2.5. Scatterplot of Factor 1 (Tree Cover) scores against Factor 2 (Coniferous Shrub Cover + Richness) scores, by buffer type.  74  Figure 2.6. Relationship between riparian buffer width and Tree Cover (Factor 1 scores + 2)  75  Figure 2.7. Winter wren activity budget. Proportion of time spent (a) engaged in foraging activities, and (b) foraging on different substrate types 76  vi  Figure 2.8. Yellow-rumped warbler activity budget. Proportion of time spent (a) engaged in foraging activities, (b) foraging on different substrate types, and (c) foraging in different vegetation layers  77  Figure 2.9. Golden-crowned kinglet activity budget. Proportion of time spent (a) engaged in foraging activities, (b) foraging on different substrate types, and (c) foraging in different vegetation layers  78  Figure 2.10. Townsend's warbler activity budget. Proportion of time spent (a) engaged in foraging activities, (b) foraging on different substrate types, and (c) foraging in different vegetation layers  79  Figure 2.11. Mean distance moved (± 1 SE) in directions relative to stream by (a) winter wrens, and (b) yellow-rumped warblers -.  80  Figure 2.12. Mean distance moved (± 1 SE) in directions relative to stream by (a) golden-crowned kinglets, and (b) Townsend's warblers  81  ACKNOWLEDGEMENTS  I thank my academic supervisor, Cindy Prescott, and my research supervisor, Tom Dickinson, for giving me the opportunity to conduct this study. To Cindy, thank-you for accepting a wildlife person into your forest nutrition flock - without your help and advice, I would not have started this project. A n d to T o m , thank-you for your never-ending support, encouragement and patience - most o f all, thanks for the beers, the laughs and the inspiration. M y committee members, Jamie Smith and John Richardson, provided excellent suggestions and feedback throughout my research. They have improved the final product immeasurably. I also thank the professors, students and staff who provided me with much-appreciated advice on statistics, G I S , experimental design and thesis writing: Lisa Zabek, Christie Staudhammer, V a l Lemay, Scott Harrison, Ralph Wells, Kathy Martin and Susan Paczek. B i g thanks to my field assistants for their hard work, patience and good humor: Ralph Heinrichs, Stu Brooks, Andrew Preston, and especially Jamie Piccin. Thank-you to Donna Falat and Dave Caswell, my Kamloops friends who volunteered their time and offered their support. I also thank the British Columbia Conservation Foundation for their logistical and administrative support, Sasha Lees for her excellent lessons in A r c V i e w , and D o n Brimacombe for his many hours o f help in T F L #35 and at the Weyerhaeuser office. Funding for the field portion of this research was provided by Forest Renewal B . C , Weyerhaeuser Canada Ltd., Mountain Equipment Co-op and the B . C . Environmental Youth Team. Weyerhaeuser also provided equipment and other inkind support. I am also extremely thankful for the personal financial support that I received through the Natural Sciences and Engineering Research Council o f Canada ( N S E R C ) and the B . C . Science Council. Finally, for all my friends and family, you w i l l never again have to ask the question "So how's the thesis coming?". To my wonderful U B C friends - K e l l y Squires, Susan Paczek, K y m Welstead, Christie Staudhammer, Stephanie Melles and Jennie Christianson - thanks for all the support, laughter, fun and good memories. To the Soundproof crew, thanks for showing me that there is life outside school. A n d to my mom, my dad and Dermot.. .thank-you for all you have given me.  INTRODUCTION In North America, songbirds face habitat loss and fragmentation through a variety o f human activities, namely deforestation, land conversion and urbanization (Terborgh 1989). Habitat loss and fragmentation can have several deleterious consequences, including an overall reduction in habitat area, increased isolation o f habitat patches, greater edge to interior ratio and decreased habitat quality (Askins et al. 1987, Wiens 1994, Faaborg et al. 1995). Area and isolation effects can result in the loss o f particular species, lowered species diversity and fewer recolonization events following local extinctions (Ambuel and Temple 1983, Freemark and Collins 1992). Edge effects, while traditionally viewed as beneficial to game species (Harris 1984, Hunter 1990), can have negative effects on non-game species in the form o f increased predation, parasitism and competition (Gates and Gysel 1978, Brittingham and Temple 1983, Simberloff 1994, Vander Haegen and Degraaf 1996, Brand and George 2000; but see Haskell 1995, Zanette 2000, Davidson and Knight 2001). A s forested areas become increasingly fragmented, the number and abundance o f forestassociated songbird species is reduced in patches (Faaborg et al. 1995, Austen et al. 2001) and along edges (Brand and George 2001). A s well, forest birds in smaller fragments may be negatively affected by increased competition with edge and openhabitat species. In British Columbia (B.C.), forest harvesting comprises a major segment of the provincial economy and occurs in much o f the province. In south-central B . C . , many lower elevation forests have been logged and harvesting is now focusing on montane and subalpine forests. Songbirds breeding in these high-elevation  1  environments face harsh climates and limited resources (Sabo 1980, Sabo and Holmes 1983), and may not respond well to reduced habitat and resource availability. Conversely, high-elevation forests in the B . C . interior have been historically disturbed by natural forces, including stand-initiating fire and insect outbreaks (Province o f B . C . 1995a). The songbird species inhabiting these forests have adapted to natural disturbance and thus may be able to adjust to forest harvesting effects (Bunnell 1995). In high-elevation forests, stream riparian ecosystems are productive areas for tree growth and are frequently subject to timber extraction. The Riparian Area Management Guidelines are used to regulate harvesting in these ecosystems within B . C . (Province of B . C . 1995b). Streams have been classified on the basis o f average channel width, fish presence and use as a community watershed. The retention o f Riparian Management Areas ( R M A s ) is required along streams in harvested sites. These areas are composed of two sections: the riparian reserve zone and the riparian management zone. The reserve zone acts as a buffer, within which harvesting is prohibited. Harvesting is permitted in the management zone, but may be modified to meet specific objectives. The width o f the R M A is specific to the stream class; all streams are assigned a management zone, but only those possessing fish or community watershed values are assigned a reserve zone. Headwater streams, generally small and rarely fish-bearing, receive little or no protection under the current guidelines and may be at particular risk o f riparian habitat degradation. In high-elevation forests, these riparian areas are used by bird species associated with both riparian and mature forest (Kinley and Newhouse 1997,  2  Wiebe and Martin 1998). Riparian buffers along streams at high elevations may not possess the same degree of avian richness and diversity as found along lower elevation streams, but they are important habitat in their own right (Wiens 1994) and can provide additional mature forest to songbirds breeding in a fragmented landscape. However, the retention of buffers along these small streams poses a trade-off between economics and conservation. Maintaining buffers on every headwater stream would provide abundant habitat to songbirds and other vertebrates, but the harvesting constraints would be excessive. Therefore, it is important to determine the value of streamside riparian buffers to songbirds, and in particular, to which species. This thesis examines the influence of riparian management strategies on songbirds breeding in a montane forest within south-central B.C. The primary goal of this project was to determine if riparian buffers provide habitat for forest songbirds in general, and specifically, whether or not wide riparian buffers provide habitat to a disproportionately greater number of species and individuals than do narrow buffers. The project objectives were twofold: (1) measure the effects of riparian buffer width on the community composition and density of breeding songbirds, and (2) describe the habitat use and associations of particular forest-associated songbird species. The first objective is addressed in Chapter 1, "The effect of riparian buffer width on breeding songbirds". The second objective is addressed in Chapter 2, "Vegetation characteristics of riparian buffers and the use of riparian buffers by foraging forest songbirds".  3  C H A P T E R 1. The effect of riparian buffer width on breeding songbirds. INTRODUCTION Some songbird species avoid small forest patches and are present only within large tracts of forest. Other species, although present in small forest patches, occur at greater density in large patches. In a study of songbird habitat associations, Whitcomb et al. (1981) classified forest songbirds as (1) forest-interior specialists, (2) interior-edge generalists, (3) edge species, and (4) field-edge species. The term 'forest-interior specialist' has since been linked to 'edge avoidance' and 'areasensitivity' in songbirds. Edge avoidance describes forest birds that nest only within ' forest interiors and tend to avoid edge habitats (Whitcomb et al. 1981); areasensitivity refers to those forest species that are uncommon or absent in smaller forest fragments (Freemark and Collins 1992). A number of habitat fragmentation studies have demonstrated that forest area is a principal determinant of songbird community composition (Askins et al. 1987, Freemark and Collins 1992, Austen et al. 2001). Freemark and Merriam (1986) concluded that forest size was the most important variable influencing the presence of forest-interior birds, and that interior species were found infrequently in small forest patches. Villard et al. (1999) determined that forest cover was a significant predictor of species presence, and Trzcinski et al. (1999) found that forest cover had a greater effect than forest fragmentation on species presence. Forest fragments do not possess a random assemblage of species, but rather contain a subset of those species found in continuous forests (Ambuel and Temple 1983, Blake and Karr 1987, Hobson and Bayne2000).  4  Although forest area is a strong predictor of species presence in forest fragments, the actual mechanisms causing the disappearance of certain species remain uncertain. The fact that Neotropical migrants are absent from forest patches much larger than their average territory size points to some qualitative change in the habitat (Faaborg et al. 1995). Area-dependent changes in the forest environment, such as edge effects, have been hypothesized as a key cause. Edge effects are those ecological characteristics associated with habitat edges that influence the species living there (Harris 1988). Forest edges are prone to altered microclimates and wind regimes, leading to changes in habitat features like vegetation species composition and structure (Chen et al. 1992). Because anthropogenic forest edges are generally an abrupt junction of forested and open habitats, they may also experience a greater level of biotic interaction. For forest songbirds, interactions include an increased number of open-habitat and/or edge species (Schieck et al. 1995, Hobson and Bayne 2000, Pearson and Manuwal 2001), increased rates of nest predation (Gates and Gysel 1978, Donovan et al. 1997, Burke and Nol 2000) and nest parasitism (Brittingham and Temple 1983, Robinson 1992, Robinson et al. 1995), and lowered food supply in small fragments (Zanette and Jenkins 2000) Stream riparian buffers (i.e., forest retained along a stream after harvesting or other land use activity) essentially represent long fragments of forest, particularly when harvesting has occurred on both sides of a stream. Riparian buffers are primarily retained for the preservation of water quality and other aquatic values, but they also provide habitat for terrestrial vertebrates confronted by an increasingly fragmented landscape. Research has indicated that riparian buffers can provide  5  habitat to a segment of the songbird community, but that long and narrow buffers may exclude certain edge avoiding and area sensitive species. Studies of mixed-wood forests in the east and northeastern United States have found that wide riparian buffers are required to maintain songbird assemblages similar to those of undisturbed forests. Croonquist and Brooks (1993) examined bird communities in stream riparian corridors in central Pennsylvania and found that buffers greater than 25 m in width were necessary to maintain disturbance-sensitive species. In Maryland and Delaware, Keller et al. (1993) discovered that Neotropical migrants were primarily present only in buffers wider than 100 m. Both Thurmond et al. (1995) in Georgia and Meiklejohn and Hughes (1999) in Vermont determined that while forest-interior specialists were abundant (# individuals/transect) in undisturbed forests, they were significantly less abundant in riparian buffers ranging from 50-164 m in width. Research has yielded similar results in the conifer-dominated forests of eastern Canada and the Pacific Northwest. In Newfoundland, Whitaker and Montevecchi (1999) found that interior forest species were rare even in the widest lakeside buffers (40-50 m) when compared with local interior forest habitat. In Quebec, a study of streamside riparian buffers by Darveau et al. (1995) determined that 60-m wide strips were required to prevent the loss of most forest songbirds. Hagar (1999), comparing riparian buffer widths in western Oregon, found that buffers >40 m wide were likely to provide the most benefit for forest songbirds. Within B.C., two studies in mid to high-elevation forests indicated that songbird density (detections/ha) increased with buffer width (Kinley and Newhouse 1997), and that  6  forest strips <38 m wide were unsuitable for five forest songbird species (boreal chickadee (Poecile hudsonica), brown creeper (Certhia Americana),  Swainson's  thrush (Catharus ustulatus), Townsend's warbler (Dendroica townsendi), winter wven(Troglodytes troglodytes)) (Gyug 1996). This chapter examines how the width of riparian buffers, retained after harvesting, affects songbirds breeding in a high-elevation coniferous forest. My objectives were to: (1) measure changes in the songbird community (richness, diversity, similarity) across a range of buffer widths, and (2) quantify variation in songbird territory density as a function of buffer width. I hypothesized that as buffer width increased, the resulting increase in the amount of forest habitat and the number of niches available would effect changes in the songbird community. I predicted that the number of forest songbird species would increase with buffer width while the number of open-habitat and generalist species would decrease. Species classifications are given in Table 1.2 and Appendix 2. I also predicted that wide buffers would possess a songbird assemblage most similar to that of the unharvested forest, and that the similarity between treatments and unharvested forest would decrease with decreasing buffer width. Finally, I predicted that increases in buffer width would result in higher densities (# territories/ha) of forest songbirds and lower densities of open-habitat and generalist songbirds.  7  METHODS Study Area and Experimental Design The study was conducted within Tree Farm License ( T F L ) #35, a 30,000 ha area o f forested land currently licensed to Weyerhaeuser Canada Ltd. T F L #35 is located in the Kamloops Forest Region and is approximately 30 k m north of Kamloops, British Columbia (Figure 1.1). It occurs in the Thompson-Okanagan Plateau Ecoregion, an area o f the South Interior Ecoprovince that consists o f a broad plateau with low elevation basins. This region has mean annual temperatures and precipitation ranging from 1.8°C and 1900 m m in high elevations to 10°C and 330 m m in low elevations, making it one of the driest and warmest climates in B . C . (Meidinger and Pojar 1991). Experimental units ranged in elevation from 1100 m to 1600 m and were characterized by east-southeast facing aspects (Table 1.1). The matrix surrounding each unit was composed o f coniferous forest interspersed with clearcut patches of various ages. Forest stands contained subalpine fir (Abies lasiocarpa (Hook.) N u t t ) , hybrid white spruce (Picea engelmannii Parry x glauca Voss) and lodgepole pine (Pinus contorta var. latifolia Engelm.), with lesser amounts o f Douglas-fir (Pseudotsuga menziesii (Mirbel) Franco). The average age o f the stands surrounding the experimental units ranged from 101 to 250 years. The study focused on two biogeoclimatic zones: the mid-elevation Montane Spruce ( M S ) zone and the high-elevation Engelmann Spruce-Subalpine F i r (ESSF) zone. Specifically, experimental units were located in the Montane Spruce, dry mild variant 2 (MSdm2) and the Engelmann Spruce-Subalpine Fir, dry cold variant 2  8  (ESSFdc2). The M S d m 2 is characterized by cold winters and moderately short, warm summers. Climax stands are composed o f subalpine fir and hybrid white spruce with an understory containing falsebox (Pachistima myrsinites (Pursh) R a f ) , black huckleberry (Vaccinium membranaceum Dougl.) and grouseberry (Vaccinium scoparium Leiberg) (Lloyd et al. 1990). The ESSFdc2 is typified by long, cold winters and short, cool summers. Engelmann spruce (Picea engelmannii Parry) and subalpine fir comprise the canopy in climax stands; the understory is dominated by white-flowered rhododendron (Rhododendron albiflorum Hook), black huckleberry and Sitka valerian (Valeriana sitchensis Bong.) (Lloyd et al. 1990). This was a retrospective study that took advantage o f existing riparian buffers. In 1998,1 located 10 experimental units throughout T F L #35 (Figure 1.2). I selected study sites that were similar with respect to stream size, clearcut age, surrounding forest type and consistency o f riparian buffer width. Sample size was constrained by the availability o f appropriate sites; treatments and controls were not selected randomly and were only replicated twice. Due to the uniqueness o f each site, a nested experimental design was used (Hicks 1993), with replicates nested within treatments. Treatments consisted of four different buffer widths, defined as "Very Narrow" (2-3 m), "Narrow" (11-15 m), " M e d i u m " (30-34 m) and " W i d e " (57-69 m) (Table 1.1) (Figure 1.3). Treated sites were located along small streams (< 5 m wide) with recent clearcut harvesting (< 4 yr) on one side and continuous forest on the other side. Control sites were placed along small streams (< 5 m wide) with > 200 m o f unharvested forest on both sides of the stream. A l l sites were surrounded by mature  9  forest and were at least 200 m from the next clearcut. Treated sites had at least one road leading into the study block; roads were absent from control sites.  Breeding Bird Sampling Spot-mapping was used to survey breeding songbirds and determine territory density and location. This method involves a series of repeat visits to well-marked grids, within which the locations of all songbirds are recorded. At the end of the breeding season, location data are used to delineate territory boundaries for different individuals (Bibby et al. 1992). I established a 4.5-ha grid at each site; this was less than the recommended area of 10 ha (Svensson and Taylor 1970) because of cutblock size limitations. Each grid was approximately 300 m long and 150 m wide, oriented lengthwise along the stream. Grid lines were flagged every 25 m; flagged markers were then georeferenced using a Global Positioning System (GPS) and plotted at a scale of 1:1500. I used these grid maps as census maps for recording all songbird observations. This permitted each bird observation to be digitized, entered into a Geographic Information System (GIS) database and assigned a georeference coordinate. Spot-map surveys were carried out from 30 June - 15 July in 1998, and from 01 June - 13 July in 1999. Because of the time required to locate study sites and construct grids in 1998, only one survey was completed at each site and I was unable to delineate territories for density estimation . In 1999, sites were surveyed 7 times, meeting the minimum number of visits required by spot-mapping protocols (Svensson and Taylor 1970).  10  Two observers conducted the one survey in 1998, and four observers conducted the first four survey sessions in 1999. The remaining three sessions in 1999 were divided between two observers. Observers were trained over several days to standardize data collection. A l l observers surveyed the same plot over the same period and then compared results. B y the end o f the training period, observers were generating similar survey results. A l l surveys were completed between 0500 and 0900 hrs P S T , and observers were rotated between sites to minimize observer bias. Survey start points were also rotated to prevent time-of-day bias. Surveys were not conducted in heavy rain or snow.  Territory Delineation with GIS I used A r c V i e w 3.1 (ESRI Inc. 1998) to delineate songbird territories from the 1999 survey data. A l l bird observations at each site and across all sessions were digitized and linked with a corresponding database. To aid in defining edge territories, the digital data included those individuals heard or seen up to 100 m beyond the grid boundaries. Database queries were used to select observations based on species, resulting in a separate map for each songbird species at each site, colorcoded by session date. Countersinging and moving individuals were coded separately to help differentiate territorial males. Songbird territories were delineated using criteria adapted from Svensson and Taylor (1970), Manuwal and Carey (1991) and Bibby et al. (1992). Territories were outlined using straight lines to connect the outer points o f observations (Worton 1987) (Appendix 1). The effective number o f visits was based on when each species was first observed on the study grids. For most species, this ranged from 5 to 7 visits;  11  accordingly, a minimum of 3 registrations of a singing male, from at least 3 separate sessions, was required to define a territory. Because of the short breeding season, records in a territory were an average of 6 days apart, not the 10 days recommended byBibby et al. (1992). I used Arc View to calculate the area of all territories inside and on the edge of each study grid. For each species, the number of territories in a grid was calculated by summing the number of complete territories and the proportion of each edge territory falling within the grid boundaries. The number of territories present in the cutblocks and buffers was calculated in the same way. I also summed territories across all species to yield the total number of territories for each grid, clearcut and buffer. I then divided the total number of territories for each species, and for all species combined, by the grid area (hectares) to calculate density (# territories per hectare) for each grid. Density values were also calculated for each cutblock and buffer by dividing the number of territories in each habitat by the cutblock area (ha) and the buffer area (ha), respectively.  Data Analysis Songbird Community Characteristics I calculated species richness, species diversity and community similarity values for each treatment. Species richness was simply the number of species with territories in each grid. Species diversity was calculated using both the ShannonWiener and Simpson diversity indices. These indices are sensitive to the rare and dominant species in a community, respectively (Krebs 1989). The Bray-Curtis index was calculated to evaluate the similarity of songbird communities across treatments  12  (Krebs 1989, Brower et al. 1990). This index o f similarity is recommended for situations where species diversity is low and sample size is small (Krebs 1989). A l l indices were determined using the density o f territories per hectare. I used regression analysis to assess the relationship between riparian buffer width and each of total species richness, buffer species richness and species diversity (SigmaPlot SPSS Inc. 2001). A l l variables were tested for a normal distribution and equal variance; where necessary, variables were transformed using either the squareroot or the logio transformation and re-tested for normality. Scatterplots o f each variable vs. buffer width were used to determine the most appropriate regression model (if any) for testing (Zar 1996). A significance level o f a = 0.05 was used for all tests. I assessed the fit o f each model with the coefficient o f determination (r ), the 2  standard error of the estimate, the overall significance o f the model and the significance of each parameter in the model (Zar 1996). From examining species composition in the riparian buffers and the clearcuts, it was apparent that species fell naturally into broad habitat associations. I characterized these associations as forest (unharvested forest and buffers), open habitat (clearcuts) and general (clearcuts and buffers). Numerous researchers have also classified songbirds according to their broad habitat associations (Freemark and Collins 1992, Darveau et al. 1995, Gyug 1996, Whitaker and Montevecchi 1997, Davis et al. 1999, Hagar 1999, Whitaker and Montevecchi 1999). I used these classifications, in combination with species-habitat associations observed in my study area, to categorize species as either forest, open-habitat or generalist (Table 1.2, Appendix 2).  13  Songbird Density At each site, I calculated the total density (# territories/ha) of all species in the entire grid, and in the buffer and clearcut. I also calculated the density of all forest and open-habitat species separately in the entire grid, buffer and clearcut. Finally, I calculated the density of each breeding species in each grid, buffer and clearcut. Only common species (i.e., those with territories on > 3 study grids) were selected for further analyses. I used the Shapiro-Wilks goodness of fit procedure to test all density data for a normal distribution prior to statistical analysis (SPSS Inc. 1999). Where necessary, variables were either square-root or logio transformed and retested. Scatterplots were used to visually assess relationships between buffer width and the different density variables. I used regression analysis to examine the effect of riparian buffer width on territory density within the grids (SigmaPlot SPSS Inc. 2001). Territory density was regressed against buffer width for all species combined, for forest and open-habitat species separately, and for the five most common species: dark-eyed junco (Junco hyemalis), golden-crowned kinglet (Regulus satrapa), Townsend's warbler, winter wren and yellow-rumped warbler (Dendroica  coronata).  Model selection was based on the coefficient of determination (r ), the standard error of the estimate, the overall significance of the model and the significance of each parameter in the model (Zar 1996). When different models generated similar significance values for the same variable, the most appropriate model was chosen based on the sum-of-squares technique (Hilborn and Mangel 1997) and the principle of parsimony (Burnham and Anderson 1998). Residuals were  14  plotted against predicted values as a final means of judging the fit of the model to the data (SPSS Applications Guide 1998). Non-parametric analysis was performed on density data that could not be normalized through transformation (SPSS Inc. 1999). I used the Kruskal-Wallis oneway analysis of variance, with a = 0.05, to assess the effect of riparian buffer width on the density of seven songbird species: American robin (Turdus migratorius), chipping sparrow {Spizella passerind), Lincoln's sparrow (Melospiza lincolnii), mountain bluebird (Sialia currucoides), Hammond's flycatcher (Empidonax hammondii), Swainson's thrush and varied thrush (Ixoreus naevius) (Zar 1996).  15  RESULTS Community Characteristics Forty-six bird species were recorded within the study grids during the 1999 spot-map surveys (Appendix 2). Thirty-six species were songbirds; the remainder consisted of three raptors, three woodpeckers, two corvids, one shorebird and one species of grouse. I considered songbirds as those species occurring within the Order Passeriformes, but excluded the Family Corvidae, due to their large territory sizes and lack of conspicuous territorial behaviour. Of the twenty songbird species that held territories within the grids (i.e., were recorded on > 3 separate surveys), I classified nine as forest, eight as open-habitat and three as generalist (Table 1.2, Appendix 2). Of these, twelve species were present on > 3 grids and were considered common (Appendix 2). The Bray-Curtis community similarity index described the similarity in species composition among treatments (Table 1.3). Overall similarity was high (> 0.595); the greatest similarity occurred between the very narrow, narrow and medium treatments, and between the wide treatments and the unharvested forest. The greatest dissimilarity occurred between the wide and very narrow/narrow treatments, and between the unharvested forest and very narrow/narrow treatments. Species richness within the grids was low at all study sites (Figure 1.4a), indicating a species-poor songbird community in general. The greatest richness occurred in the wide treatments and the lowest occurred in the yery narrow treatments and the unharvested forest, although overall differences were small. Species diversity was also low across all treatments (Figure 1.4b). Despite measuring different aspects  16  of the community (see Methods), both the Shannon-Wiener and the Simpson indices showed the same pattern - that of maximum diversity in the wide treatments and minimum diversity in the unharvested forest. Regression analysis did not reveal a significant relationship between riparian buffer width and either of total species richness, Shannon-Wiener diversity index or Simpson diversity index. However, there was a significant positive relationship between riparian buffer width and buffer species richness (Figure 1.5). The total number of species in the buffer grew rapidly with small increases in buffer width, and then began to plateau as buffer habitat occupied greater portions of the survey area. A power model yielded the best fit (P<0.001), explaining 79% of the variation in the data.  Songbird Density Regression analysis indicated a significant positive relationship between riparian buffer width and the total density (# territories/ha) of all species within the study grids (Figure 1.6a). The greatest increase in total density occurred between the narrow and wide buffers (Table 1.2). The logarithmic model provided the best fit and explained 59% of the variation in the data. Riparian buffer width also had a significant effect on the density (# territories/ha) of forest species within the study grids (Figure 1.6b). Density increased steadily from the very narrow to the wide buffers, after which the rate of increase slowed. The logarithmic model provided a simple, yet highly significant, description of the relationship and explained 88% of the variation in the data. Increasing buffer width did not have a significant effect on the density (# territories/ha) of open-habitat species within the study grids. Only in  17  the controls, where open habitat was completely lacking, were all open-associated species absent (Table 1.2). I conducted regression analysis on the five most commonly occurring songbird species within the study grids (dark-eyed junco, golden-crowned kinglet, Townsend's warbler, winter wren and yellow-rumped warbler). Four species had significantly greater densities in wider buffers: golden-crowned kinglet, Townsend's warbler, winter wren and yellow-rumped warbler. There was no clear relationship between buffer width and dark-eyed junco density, although juncos were absent from the unharvested forest (Table 1.2). Golden-crowned kinglet and Townsend's warbler densities displayed a similar relationship with buffer width (Figure 1.7 a, b). Density rose rapidly with small increases in width, and then slowed as buffers became wider and reached a maximum in the unharvested forest. In both cases, the logarithmic model yielded the best fit, explaining 89% of the variation in the kinglet data, and 85% of the variation in the warbler data. Although the maximum densities of golden-crowned kinglets and Townsend's warblers were similar, the slope of the regression model was slightly higher for kinglets, indicating a stronger response to small increases in buffer width. Winter wrens and yellow-rumped warblers also showed significant responses to increasing buffer width, although the models failed to account for over half of the variation in the data. Winter wren density increased with width to a maximum in the unharvested forest (Figure 1.7c). A linear model provided the best fit and explained 43% of variability in the data. The density of yellow-rumped warbler territories also rose with increasing buffer width, but reached a peak in the wide buffers, then  18  declined again in the unharvested forest (Figure 1.7d). Although a quadratic model was significant only at a = 0.10, it accounted for 48% of variation and provided the best fit to the yellow-rumped warbler data. Non-parametric Kruskal-Wallis tests were conducted on the seven commonly occurring songbird species that did not exhibit normally distributed territory densities within the study grids (Table 1.4). At a = 0.05, riparian buffer width did not have a significant effect on the density of any of the seven species. However, if the tests are considered exploratory and a = 0.10 is used, buffer width had a significant negative relationship with the Lincoln's sparrow, and a significant positive relationship with the Swainson's thrush and the varied thrush.  19  DISCUSSION  Width Effects on Community Characteristics The study area possessed a fairly simple songbird community, dominated by a few common species. Species number was similar to that observed in other highelevation and boreal studies (Darveau et al. 1995, Schieck et al. 1995, Kinley and Newhouse 1997, Schmeigelow et al. 1997). The low richness and equitability, typical of high-elevation and boreal forests, is likely the result of the reduced productivity and short breeding seasons found in these environments (Able and Noon 1976, Sabo and Holmes 1983, Osborne and Green 1992, Folkard and Smith 1995). The assemblage of songbirds breeding in the study grids was characteristic of that found in montane forests throughout British Columbia (Gyug 1996, Kinley and Newhouse 1997, Wiebe and Martin 1997). Although numerous songbirds were considered 'rare' for the study area (i.e., recorded only once or twice), none of these species was provincially listed as threatened (blue-listed) or endangered (red-listed). Many of the species described as rare in this study (e.g., cedar waxwing (Bombycilla cedrorum), warbling vireo (Vireo gilvus), brown creeper) are actually common at the regional level. The forest species that were absent from the narrower buffers (e.g., Swainson's thrush, varied thrush, Hammond's flycatcher) are also considered regionally common. Species richness and diversity were higher in the buffers than in the unharvested forest, and both values increased with buffer width. The increase in richness and diversity with buffer width may have been due to the influence of habitat heterogeneity (Triquet et al. 1990, Dickson et al. 1995, Gyug 1996, Hagar 1999, Meiklejohn and Hughes 1999). Grids containing very narrow and narrow buffers  20  were primarily composed of clearcut habitat and were inhabited mostly by openhabitat and generalist species. Grids containing medium and wide buffers incorporated a relatively large amount of forest adjacent to open habitat; accordingly, these sites had a mixture of generalist, open-habitat and forest species. The low richness and diversity recorded in the unharvested forest reflects the loss of species associated with clearcuts, forest openings and edges. Increasing riparian buffer width had a positive effect on the number of species occurring in the buffers. Despite the presence of continuous forest on the opposite side of the stream, the richness of forest songbird species increased significantly with incremental changes in the amount of habitat. The observed effect of buffer width on buffer species richness implies that certain forest species may have been areasensitive and/or edge-sensitive. Keller (1993) found a similar pattern for Neotropical migrants in streamside buffers and attributed it to the presence of various areasensitive species (e.g., Acadian flycatcher [Empidonax virescens], red-eyed vireo [Vireo olivaced]) in wider buffers. The relationship between buffer width and species richness may also have been due to the increased availability of upland habitat in the medium and wide buffers, which attracted forest species not commonly associated with riparian habitats (Wiebe and Martin 1997).  Width Effects on Songbird Density Total Density The total density of all species and the density of forest species both showed significant responses to small changes in the amount of buffer habitat. The similarity of these patterns suggests that forest bird species were largely responsible for the  21  observed relationship between total density and buffer width. As with buffer species richness, density increased with buffer width despite the presence of contiguous forest on the opposite side of each stream, indicating the possible sensitivity of several forest species to small-scale changes in habitat availability. Similar results were yielded by two other studies with riparian buffers on one side of the stream only. Kinley and Newhouse (1997) found that the total density of all species differed significantly among buffer widths, with higher densities in wider buffers.  Similarly,  Darveau et al. (1995) found that forest songbirds used all remnant forest buffers, but that their densities significantly declined in narrow buffers (20 m) over a three-year period. The density of open-habitat species did not change significantly with increasing riparian buffer width. Unlike their forest-dwelling counterparts, openhabitat species were not sensitive to small changes in the availability of their habitat. Although the ratio of buffer-to-clearcut within the study grids increased with increasing buffer width, the study grids were still surrounded by open habitat. These results suggest that the density of open-habitat species was primarily influenced by the availability of clearcut habitat on a larger scale.  Species-Specific Density Most open-habitat species occurred at relatively low densities in the study grids, and all but the dark-eyed junco and the MacGillivray's warbler reached peak density in the very narrow (2-3 m) and narrow (11-15 m) buffer sites. The dark-eyed junco was the most abundant open-habitat species and was present in relatively high densities across all treatments. Juncos were frequently observed utilizing the forest  22  edge and residual patches of trees within the cutblocks for singing. Although all open-habitat species were common in the early serai habitat provided by the clearcuts, they were absent from the undisturbed forest. This contrasts with Davis et al. (1999), who found that species such as the American robin, chipping sparrow, Lincoln's sparrow and dark-eyed junco occurred in both early-seral and mid- to lateseral habitats in high-elevation forests.  The lack of open-habitat species in the  unharvested forests of this study may reflect a difference in structure (such as greater canopy closure and fewer canopy gaps) or a preponderance of open habitat available throughout the study area. Generalist species occurred with the greatest density where both open and forest habitats were present, and appeared to benefit most from the juxtaposition of clearcuts and riparian buffers > 30 m. Wilson's warblers were present only in medium (30-34 m) and wide (57-69 m) buffer sites, and yellow-rumped warblers reached maximum density in the wide buffer sites. Territories of both species included buffer and clearcut habitat, and individuals were frequently observed singing and calling from the forest edge. Yellow-rumped warblers also often used the small, residual patches of trees left in most cutblocks. In contrast, Kinley and Newhouse (1997) found the highest density of yellow-rumped warblers in medium buffers (37 m) and Whitaker and Montevecchi (1999) detected Wilson's warblers only within buffers. The density of these insectivorous species in the medium and wide buffers of my study may, in part, be a response to greater food availability. In western Newfoundland, Whitaker et al. (2000) linked the abundance of ubiquitous birds, such  23  as the yellow-rumped warbler, in riparian buffers to the elevated numbers of largebodied flying insects found along the edges of lakeshore buffers. All forest species reached their highest densities in the widest riparian buffers and the undisturbed forest. However, the absence of varied and Swainson's thrush from the medium and narrow buffers may indicate that buffers < 34 m are not sufficient to provide suitable habitat for these species. Alternatively, these species were relatively uncommon in the study sites and may have been too sparsely distributed to sample effectively. Hagar (1999) found that the varied thrush was absent from most buffers, and Kinley and Newhouse (1997) observed it only in the widest buffers (70 m). Conversely, the sensitivity of Swainson's thrush to riparian buffer width may depend on ecosystem type or forest structure. Although Swainson's thrush has been characterized as an interior specialist in eastern Canada (Whitaker and Montevecchi 1997, 1999) and width-sensitive in western Canada (Gyug 1996), both Darveau et al. (1995) and Kinley and Newhouse (1997) recorded Swainson's thrush in all buffer widths. Three forest species occurred in all riparian buffers: golden-crowned kinglet, Townsend's warbler and winter wren. The presence of these species in even the narrowest buffers suggests that, at least in a high-elevation forest, they may not be particularly sensitive to forest edges. Although they exhibited some sensitivity to area (i.e., higher densities in wider buffers), the three species did not seem to depend on interior forest habitat. Golden-crowned kinglets were recorded in all buffers, with medium buffers (30-34 m) retaining approximately 60% of the kinglet density found in the  24  unharvested forest. Increasing the width to 60 m produced only a 10% further increase in density. These findings differ from those of other studies, where goldencrowned kinglets were described as highly area-sensitive. Whitaker and Montevecchi (1999) observed that golden-crowned kinglets were absent from all riparian buffers and Darveau et al. (1995) noted that kinglet density was very low in buffers <20 m. Hagar (1999) recorded low densities in even the widest buffers (40-70 m) and suggested that golden-crowned kinglets avoid edge and/or riparian habitats. Townsend's warblers were also recorded in all buffers, although avoidance of riparian habitat was suggested by the observation that territories were only located in upslope areas in the unharvested forest. Medium buffers (30-34 m) retained 50% and wide buffers (57-69 m) retained 80% of the Townsend's warbler density found in the unharvested forest. Kinley and Newhouse (1997) noted that Townsend's warblers were present in all buffers, though at much lower densities than golden-crowned kinglets.  In the high-elevation forests of southern B.C., Gyug (1996) observed that  Townsend's warblers were absent from buffers <47 m and termed the species 'widthsensitive'. This range of results indicates that the sensitivity of Townsend's warblers to buffer width varies, perhaps in response to the changing nature of buffer habitats across the different studies. Winter wrens were recorded in all buffers and most territories spanned the creek, even in the unharvested forest. Associated with, but not dependent upon riparian habitats (McGarigal and McComb 1992, Waterhouse 1998), winter wrens located their territories primarily along streams. Because a large portion of the riparian habitat was occupied in each buffer, territories along the stream did not  25  increase in number with buffer width. Instead, the number of territories that occurred in moist, upslope areas increased. Medium buffers (30-34 m) retained 65% and wide buffers (57-69 m) retained 80 - 90% of the density found in the unharvested forest, suggesting that winter wrens were not particularly sensitive to buffer width. Whitaker and Montevecchi (1999) detected similar winter wren densities in both controls and buffers. However, Kinley and Newhouse (1997) recorded low numbers of winter wrens in all but the widest buffers (70 m), and Gyug (1996) found winter wrens to be absent from buffers <47 m. From these differing results, it seems that the winter wren, like the golden-crowned kinglet and the Townsend's warbler, exhibits a range of sensitivity to riparian buffer width.  Edge Effects Forest edges have a range of effects on a number of variables, including microclimate, vegetation structure and vertebrate species distributions (reviewed in Kremsater and Bunnell 1999). Studies have documented a wide range of edge effect distances on birds, but many indicate that the greatest effects occur within the first 50 - 100 m into a forest (Gates and Gysel 1978, Hansson 1983, Vander Haegen and Degraaf 1996, Brand and George 2000, Brand and George 2001). A number of studies report increased bird species richness and density near edges, the result of incursions by species preferring the open and often more shrubby habitat that occurs at edges and in adjacent clearcuts (reviewed in Kremsater and Bunnell 1999). Conversely, forest-dwelling species exhibit an opposite response, avoiding the altered habitat and increased risk of predation/parasitism found along edges (Kroodsma  26  1982, Brand and George 2001, Flaspohler et al. 2001, but see Campi and Mac Nally 2001). The riparian buffers in my study were not isolated strips of forest surrounded on all sides by clearcut habitat. Instead, buffers were on one side of the stream only, with continuous forest located both up and downstream, and across the stream. As a result, the stream nominally defined the width of each buffer. Small streams (i.e., < 5 m) do not likely act as barriers to most forest songbirds. Hence, the study grids could have been measuring the response of songbirds to increasing distance from the forest/clearcut edge, as well as to changing amounts of clearcut and forest habitat within each grid. As buffer width decreased, the relative amount of forest within the study grid also decreased while the proportion of buffer habitat influenced by edge effects increased. The resulting patterns of forest bird richness and density were likely a combination of both factors. In assessing the response of songbirds to riparian buffer width, it is difficult to separate the effects of forest edge from the effects of forest area. However, it may be hypothesized that the close proximity to edge in riparian buffers results in a lower richness and density of forest songbirds than would occur in forest interior conditions. If edge effects decrease as buffer width and distance from the edge increase, and the greatest influence of edge on songbirds occurs within 50 m of an edge, then narrower buffers are dominated by edge and wider buffers contain both edge and interior habitat. Consequently, narrower buffers would be predicted to possess a low number and density of forest species because they consist entirely of edge; wider buffers would be predicted to possess edge habitat with reduced numbers of forest birds, and  27  interior habitat with higher numbers of forest birds. Calculations of forest bird richness and density for each buffer in its entirety would therefore be a combination of lower numbers in the edge habitat and, for the wider buffers, higher numbers in the interior habitat. If the effect of edge could be removed, calculations of richness and density based only on interior habitat may yield higher values for the wider buffers. Although this would result in larger slope coefficients for the regressions of buffer width against buffer species richness and forest songbird density, it would not change the basic pattern of increasing richness and density with increasing buffer width. Nonetheless, the evidence for negative edge effects on songbirds is not as strong in western forests fragmented by harvesting as in eastern forests fragmented by agriculture (Kremsater and Bunnell 1999). Edges between forest and clearcut habitats are not permanent; sharp edges gradually become less distinct as clearcut habitat matures into young forest. Studies conducted in the Pacific Northwest do not indicate a strong association between forest birds and interior forest (McGarigal and McComb 1995, Schieck et al. 1995). This suggests the possibility of an alternative hypothesis, where the effect of edge is minimal, and the increase in forest bird richness and density is primarily a function of forest area (buffer width). However, the sampling design and data collected in my study do not allow the effects of forest edge and relative forest area to be easily partitioned and independently evaluated. Given the nature of the data, all that can validly be concluded is that forest birds increase in richness and density with increasing riparian buffer width. The actual mechanism(s) causing this increase - whether edge effects, area effects or a combination of both - can only be hypothetically discussed.  28  Conclusions Overall, wider riparian buffers appear to support a songbird community similar to that of unharvested forests. Results indicate that buffers 30 m wide can provide habitat for at least 50% of the forest species, at >50% of the density, found in the unharvested forest. However, buffers of at least 60 m in width are required to maintain most forest species at up to 85% of the density found in unharvested forests. Likewise, Hagar (1999), Pearson and Manuwal (2001) and Kinley and Newhouse (1997) recommend minimum buffer widths of 40 m, 45 m and 70 m, respectively. However, riparian buffers are not islands - they are adjacent to clearcut habitat that will develop quickly into young forest, then slowly into mature forest. Davies et al. (1999) indicated that in the ESSF forests of east-central British Columbia, species such as the varied and Swainson's thrush inhabited forests as young as 40 years. Thus, narrower buffers that are unsuitable for some species will not remain so for long. When making recommendations regarding riparian buffers, it is important to consider benchmarks. Recommended buffer widths are a guide for managers; they provide an idea of what species may be lost and what species may be maintained in buffers of certain widths. But if the management goal is to retain a full complement of forest songbird species in a harvested landscape, some a priori decision should be made as to how much of the community can be lost before management action is taken. One such benchmark could be loss of certain species that are indicative of ecosystem health and that are representative of the habitat needs of other species (e.g., woodpeckers represent secondary cavity nesters and indicate snag availability).  29  Within my study area, the point at which the decrease in number or density of forest species becomes greater than the loss of mature forest could be another benchmark. For example, if 50% of the forest within a defined area was removed and 80% of the forest species disappeared, the loss of species would be disproportionately high relative to the amount of forest harvested and should be cause for management concern. At my study sites, the loss of forest species was not disproportionately greater than the loss of forest. Study grids with wide buffers had approximately 57% of the forest removed and lost about 20% of the forest songbird species; study grids with narrow buffers had approximately 90% of the forest removed and lost 50% of the forest species. At the stand level, my findings do not generate serious concern or require strong management action. Nonetheless, scale is an issue; highly mobile organisms such as songbirds, and bird species with large area requirements (e.g., hawks, owls and woodpeckers), should be assessed at a broader scale. Within T F L #35, the loss of forest birds could be measured at the watershed level. At a regional scale, benchmarks could be assessed at the biogeoclimatic zone or subzone level. A disproportionately high loss of species or individuals over a large area or within certain habitat types should necessitate a change in forest planning and management.  30  > a S3  u «  i  2 g >- «  CO ON sO © © CN r<">  CN  m  © CN vo'  2; £  O0C/300O0O0 00 O0O0 00 00  —•  CO w  c  w  M  w  o  o  o  ^  W pq qq >. czi oo !>  w  ^ 2;  o  o  o  o  o  o  ^ ^  o  IS  00 00 00  cfl ©,  o  o OS  A A  os —. w  i—I CO  XI  as OS < 1 w  -O PQ  o i> c .S  c3 — i  <3  CN —i  o  b  >>  O  CN  CN  s  —< CN CD  -a T3 '•3  —i —  o a o U  CN _  o c o U  is  CO  §  ^ *  31  ON o  o  o  o  o  o  o  o  o  o o o o o o o o o d d d d d d d d d  oo  u B -a  o o o o o o o o o o o o o o o o o o o d d d d d d d d  oo <D  o-*—'Or--om(NO'o  O O  o o \ o o r - ^ O N O c N O N O  O O  OS  I O O O O O — ' C N O - ^ n  ON  —  -H  l O O C N © © — ' O N C O C O C N  d o d o  mio—• o o\ © © © o — m o o o o o  o O d o  t-^ (N —i o — o d o c~- 0\ <J~i O — i "3" *o d o d o  o o o o o o o o o r O V ^ U - l O O s © © © —' O — ' \ O O O O O O O N  oo  o o co m o o o o © do ©  o d d d d d d d d  O O — ' - - ' O O O ^ - ' O C N  d d d d d d 0 , 0 , 0 ©,  P'"*. ~ : ~ ~ : P ! ~ ! ! ! " 1 < N  l /  r r  r s  o o o o o o o o d c N  m  a o * .2 00  ©  r- —i CN © NO © o o —' o o o o o  o o — o CN d d d d  O r o N O O O O U - l ^ - N O C N © © — © © © © © © c o  cNt^O\r^ONO'*toin o o ^ j - o o o o o t ^ c6Sd>SSd>d>d'd  o r~-  O O O N O O O O i O T f O O O N o n — O O O C N O — 1 o\  O  d d d d d d d d d d  i n — ' CN T j -  d>d>ddd>cid''d'd>d>  odd©  GO  4J_ —  O O  ^r-~-*oooo^o —  O O O O O C N  ©©odd©©©©  00  CS  s  CN NO O O , 0,  O  0 O 0, d 0 0 ir> 0 d d O d  s  >  — --I  © co  ON  © -H  O  © ©  ©odd  0 0 NO NO 0 © ©©© d 0 d d  0 0 0 ©© CN 0 0 d  -•  CN O  O  © —  CO  ©  ©  ©  © -H ©—  ©  o o o © o © © o o o  --  ©0 ©0 0, d © ©  © © © -H c o © © -H O © © O © ©  ©0©©©00©00  © c o CN O — CN  0 00 NO 0 0 ©O m ©© 0, ©  © 0 0 0 ©0 0 0 0 ©d d d d O CN 0 0 0 ©0 0 0 0 d d d d d  00 CN O  0 0 d 0 0 d  > / > - 0 sO 0 0 0 0 d d d d 00 0 CN CN 0 0 CN m d d d d  00  d 00  1/-)  •—  1  CN  c <S oo (3 o O  M a  o 00 00  s  u cu  -*-t 03 -*—»  _CU  '3  a.'  C/3'  I1  cu  cu  a. e  a.  es  a  O  ©  ON  <z><sd<d<Sc>d>Sd>  o  O  ©  d o d o  — o\r-Tfooooo o—'co — o o o o r ^  oo c U -a c  ON  SI s o  cd  £ ^> ISC O  O o •S & e ID  s  cu  O o 2 « .S ca UDiJ 2 2  c  03 CU  1-  a  Z  TS 3  u o  .5 CCi  c  cu  cu u ,0  u cu  a. 60 .S -c  • — -4—1  1 "S  Uc cu C ^ = c ^ o  H  c  CO  H *  s ^-S  I  cu  o  «i  .2  o  00  B  S  "O  15  cu  B'  on  C3  T3  C cd5 S O  <D  S £ 1- o  cu •o C3  >  E o H  32  in in  ON  ON m  ro oo  o  oo ro oo ©  o u  «  ©  >o T J ON ro in  ©  \o  o  oo ON m ON m ON m ON oo i n * o ©  ©  ©  ©  o  =3 o C  0)  u c «  IT) OO CN i o OO 00 vO 0 \ O i 0\ - H O O O O CN U-> ^  ©  ©  ©  ©  ©  CN VO oo od  <r> ^ - i ^-i "3" CN CO OO ON oo oq r-~^ r-~ iri CN  ©  ©  ex  133  « 3  CP  a  l<»l C/3  CD  J! o  3 1CD  03  H  (N \D ^ ON ON in CO r-^  Figure 1.1. Location of Tree Farm License ( T F L ) 35 within the Kamloops Forest Region ( K F R ) , in the south-central portion of British Columbia.  35  Treatments C Control • Wide • Medium A Narrow * V Narrow j 1 TFL boundary Creeks  3000  3000  6000 Meters  Figure 1.2. Location of study sites within T F L #35. T F L #35 is situated approximately 30 km north of the city of Kamloops, British Columbia.  36  A.  Uncut  Wide  Medium  Narrow  V Narrow  Buffer Type B.  Uncut Shannon Diversity Simpson Diversity  Wide  Medium  Narrow  V Narrow  Buffer Type  Figure 1.4. (a) Mean species richness (± 1 SE), and (b) mean Shannon-Wiener and Simpson diversity (± 1 SE) in study grids across buffer types. For each buffer type, n = 2.  38  10  Buffer Width (m)  Figure 1.5. Relationship between riparian buffer width and species richness in the buffers only.  39  A.  3.0  Buffer Width (m) Figure 1.6. Relationship between riparian buffer width and the territory density (square-root transformed) of (a) all species combined, and (b) forest species, within the study grids.  40  A.  B.  Buffer Width (m) Figure 1.7. Relationship between riparian buffer width and the territory density (square-root transformed) of (a) golden-crowned kinglets (b) Townsend's warblers (c) winter wrens, and (d) yellow-rumped warblers, within the study grids.  41  C H A P T E R 2. Vegetation characteristics of riparian buffers and the use of riparian buffers by foraging forest songbirds. INTRODUCTION Studies o f forest fragmentation effects on songbirds have demonstrated a • strong relationship between forest area and species composition (Whitcomb et al. 1981, Blake and Karr 1984, Faaborg et al. 1995). In particular, research on forest patch sizes (Galli et al. 1976, Freemark and Collins 1992, Schmiegelow et al. 1997) and riparian buffer widths (Darveau et al. 1995, Thurmond et al. 1995, Hagar 1999) has shown that many forest-dwelling species are constrained by habitat area and are absent from small forest patches. It has been theorized that area-dependent changes in songbird communities are produced by area-dependent changes in the forest environment (Ambuel and Temple 1983, L y n c h and Whigham 1984). When harvesting occurs, adjacent residual forests experience a number o f edge effects, including significant microclimate changes (Chen et al. 1995) that can extend several hundred meters into the forest and alter vegetation structure and composition. Reduced forest area can also cause the loss o f important habitat components, such as spatial variability in plant species and forest structure (Freemark and Merriam 1986). M a n y studies have focused on the relative importance of forest area and forest habitat in determining the composition of songbird communities in fragmented landscapes. Askins et al. (1987) hypothesized that the environment (forest size, type o f vegetation) within and immediately adjacent to a forest patch was the primary determinant o f avian richness and density. Similarly, Robbins (1989), Blake and Karr (1987), and Doherty et al. (2000) proposed  42  that patch size, isolation and habitat characteristics all contributed to songbird density and distribution with varying degrees of influence. Identifying the habitat attributes with which songbirds are associated can help to predict fragmentation effects on songbirds (Ambuel and Temple 1983). M a n y avian studies have investigated species-habitat relationships to help determine the impacts o f human disturbance (Anderson and Shugart 1974, Clark et al. 1983, Hansen et al. 1995, Hagar et al. 1996). Multivariate and regression techniques have been applied to examine relationships between bird richness or abundance and a range o f vegetation variables. Significant associations have been found with such attributes as tree species richness and canopy cover (James and Warner 1982), canopy volume, tree density and understory cover (Clark et al. 1983), and the density o f small shrubs (Swift et al. 1984). Multiple regression methods have also been used to develop species-specific habitat models. Models derived for western forests have correlated the abundance of common forest-dwelling songbirds with tree density, tree diameter and tree species (Morrison et al. 1987, Hansen et al. 1995, Hagar et al. 1996). Songbird habitat models may help to identify some important habitat attributes, but they have often met with limited success when validated or tested against independent data sets (Stauffer and Best 1986, Morrison et al. 1987). A n alternative means o f examining songbird habitat use is to observe the behaviour o f individual birds in the field. Behaviour sampling can yield information on speciesspecific patterns o f utilization and preferences for certain vegetation attributes (MacArthur 1958, Robinson and Holmes 1982, Sabo and Holmes 1983, Osborne and Green 1992). Observations of songbird behaviour have also been used to examine  43  how habitat alteration affects foraging behaviour and community structure (Mauer and Whitmore 1981, Mahony et al. 1997, Brotons et al. 1998). Behaviour sampling can therefore provide researchers with additional information on patterns of habitat use by songbirds in response to forest harvesting. This chapter examines the forest attributes with which songbirds are associated, and the effects o f riparian buffer width on those attributes.  M y objectives  were to: (1) measure changes in forest vegetation across buffers o f different widths, (2) determine i f the density of forest songbirds is related to changes in buffer vegetation, (3) describe foraging habitat use by four songbird species, and (4) quantify the influence o f riparian buffer width on the movements o f four songbird species. I hypothesized that decreasing riparian buffer width would result in various edge effects, including reduced tree density and cover o f tall shrubs (due to harvesting disturbance and windthrow), and increased cover o f low shrubs and herbaceous plants (due to increased light levels). I predicted that these changes in forest structure would correlate with forest bird density independently o f buffer width. Finally, I predicted that narrower riparian buffers would restrict the movements o f forest birds and result in activities that were concentrated along the stream, rather than in the upland.  44  METHODS Habitat Sampling Habitat mapping and vegetation sampling were used to describe each study site. I mapped large-scale features to make broad comparisons between habitat types and the distribution of species territories. I used vegetation sampling to measure the finer-scale attributes of habitat structure and composition. Habitat mapping was conducted in July and August 1998. Major habitat features were recorded onto survey maps (1:1500) and included riparian buffers, root masses, wildlife trees, slash and blowdown piles, and steep slopes. I then digitized these habitat features into ArcView 3.1 (ESRI Inc. 1998) and added them to the map of each study grid. Within the GIS, I overlaid the habitat maps with territory locations to display territory position relative to the different habitat features. Digital Terrain Resource Information Management (TRIM) maps and forest cover data were provided by Weyerhaeuser Canada. These included vegetation, topographical and harvest information for areas surrounding the study sites. Layering this graphical information onto the study grid maps allowed the comparison of study sites with respect to elevation and adjacent stand features. Vegetation was sampled in July and August 1998. I systematically placed 12 circular vegetation plots (0.03 ha) throughout each spot-map grid, at increasing distances from the stream: 10 m, 40 m, 70 m and 125 m (Appendix 3). Two 20-m line transects were placed within each vegetation plot, oriented at right angles to each other and intersecting at the plot center. Vegetation plots were used to measure  45  canopy attributes; line transects were used to sample understory and ground cover attributes. Vegetation plots were divided into four quadrants and the number of trees in each quadrant was tallied by species and diameter-at-breast-height (dbh) class (Noon 1980). I defined diameter classes as: 10-20 cm, >20-40 cm, >40-60 cm and >60 cm. A l l live trees >10 cm dbh and all standing dead snags >10 cm dbh and >1.5 m height were included. To estimate canopy height, a Suunto clinometer and hip chain were used to measure the height o f one tree from each D B H class present within the plot. Canopy cover along each cardinal direction from the center o f the plot was estimated with a convex densiometer (Martin et al. 1997). I used the line-intercept method (Cook and Stubbendieck 1986) to measure shrub cover by species and height class along the two line transects in each vegetation plot. Height classes were defined as: 0-1.0 m, >1.0-3.0 m, >3.0-5.0 m and >5.0 m. The distance (meters) covered by each shrub that intersected the meter tape was recorded, along with the species o f shrub and the height class. The percent cover o f each shrub species in each height class was then calculated as the total distance intersected divided by the entire transect length (40 m), multiplied by 100. I used the point-intercept method (Bonham 1989) to estimate ground cover, by cover class. Ground cover classes were defined as: herb, grass, moss, lichen, coarse woody debris, litter and bare ground. A wooden rod with a pointed metal end and a carpenter's level was dropped at 2 m intervals along both line transects. A n y cover type that was intersected by the metal point was recorded. The percent cover o f each class was then calculated as the number o f times the ground cover type was  46  intersected divided by the total number o f times the point was dropped (20), multiplied by 100.  Behaviour Sampling I chose focal individual sampling with continuous recording to sample foraging behaviour (Altmann 1974, Lehner 1987). M y first objective was to record the foraging activities of individuals to determine which foraging substrates, vegetation species and vegetation layers were used most frequently. M y second objective was to record the distances and directions moved by foraging individuals to determine how songbird movements were influenced by riparian buffer width. The winter wren and the yellow-rumped warbler were chosen as the primary focal species for behavioural observations. The winter wren, a cavity-nesting insectivore, was chosen to represent species that utilize the ground and shrub layers for foraging and/or nesting. The yellow-rumped warbler, also an insectivore, was chosen to characterize species that primarily dwell in the mid to upper canopy layers. Additional data were collected on the Townsend's warbler and the golden-crowned kinglet. Songbird behaviour was sampled from 22 June 1999 to 28-July 1999. Observations were conducted within the study grids from 0700 - 1200 hrs P S T , during the most active period o f the day. A n individual bird was followed for a maximum o f 30 minutes and foraging activities were continuously recorded with a tape recorder. There was no minimum recording time. Although sites were sampled an equal number o f times, all species were not necessarily observed at each site during each session; as a result, species observations were not distributed evenly  47  across all sites. I assumed that samples were independent, though some individual birds were likely sampled more than once at each site. For each session, data were recorded under seven categories: habitat, activity, substrate, relative position, height above ground, movement direction and distance traveled. Habitat included the forest, forest-clearcut edge, clearcut and residual tree patches within the clearcut. Activity classes included five different foraging methods - ground gleaning, bark gleaning, foliage gleaning, hover-glean and hawking (Erlich et al. 1988). Substrate referred to the item on which an activity occurred and included ground, herb, coarse woody debris, root mass, shrub, snag and tree (sapling, pole or mature). Relative position assigned the location o f a bird to low, m i d or upper shrub/trunk/canopy (where applicable), and height referred to the estimated height o f the bird above the ground while on a substrate.  Movement direction was defined  according to a set o f angles numbered relative to the stream (Figure 2.1a) and distance was the estimated number o f meters flown between substrates.  Data Analysis Forest Structure A l l vegetation variables were tested for a normal distribution using normal probability plots and the Kolmogorov-Smironov goodness o f fit procedure (SPSS Inc. 1999). A l l percent cover variables were arcsin-square-root transformed prior to analysis. I transformed additional non-normal variables using the logio transformation, then retested for normality. Vegetation data included the following variables for each plot: •  basal area (m /ha), canopy cover, tree species richness;  48  •  total shrub cover, coniferous and deciduous shrub cover in each height class and in total, conifer seedling/sapling species richness, deciduous shrub species richness, number of shrub layers;  •  cover of coarse woody debris, bare ground, litter, moss, grass and herbaceous plants. I used a one-way analysis of variance to test if riparian buffer width had an  effect on individual vegetation variables (SPSS Inc. 1999). Analysis included the continuous variables listed above and used only the data from the vegetation plots in the buffers and unharvested forest (i.e., excluded clearcut vegetation data). Variables were tested with a nested model, where replicates were nested within treatments, and a significance level of a = 0.05 was used for each test. Bonferroni multiple comparison tests were used on the variables for which riparian buffer width had a significant effect. I used factor analysis with Equamax rotation to reduce the number of vegetation variables in order to describe overall forest structure (Tabachnick and Fidell 1996). Analysis included all vegetation variables (25 in total) and used only the data from the vegetation plots in the buffers and unharvested forest. I used the correlation matrix and principal components method of extraction, and selected the number of factors based on eigenvalues (> 1.0) and scree plots. The scree plot consists of eigenvalues plotted against factors; the point at which a line drawn through the points changes slope indicates the approximate number of factors in the data. Variables with component loadings greater than ± 0.55 were selected for factor interpretation; a conservative cut-off value of 0.55 was used following the  49  recommendation of Comrey and Lee (1992). I then used regression analysis to explore the relationship between riparian buffer width and the resulting forest structure factors (SigmaPlot SPSS Inc. 2001). Each factor was regressed against buffer width separately, using a significance level of a = 0.05. I assessed the fit of each model with the coefficient of determination (r ), the standard error of the 2  estimate, the overall significance of the model and the significance of each parameter in the model (Zar 1996).  Songbird-Habitat Associations Songbird Density and Forest Structure I used partial correlations to test the relationship between forest structure and the richness and density of forest songbird species (SPSS Inc. 1999). Because previous analyses indicated that buffer width had a significant effect on both songbird richness/density (Chapter 1) and forest structure, partial correlation coefficients were calculated to test the relationship between birds and habitat while controlling for the effects of buffer width. Forest structure was represented using the four factors derived from the factor analysis. Partial correlations were computed between mean factor scores and each of forest species richness, forest species density, and the density of golden-crowned kinglets, Townsend's warblers and winter wrens. Scatterplots of mean factor scores against each songbird variable were used to assess whether or not the assumption of a linear relationship was met (Tabachnick and Fidell 1996). A significance level of a = 0.05 was used for each partial correlation.  50  Habitat Use and Movements Activity budgets were calculated for each o f the four study species, summarizing the proportion o f total time spent in each activity and on each substrate. I also summarized the proportion of time spent in each habitat type and, for canopydwelling species, the proportion o f time spent in each canopy layer. Data were pooled across all sites and both sexes; I did not examine differences i n habitat use between treatments or gender. Male wood warblers tend to forage higher in the canopy than females during the breeding season (Morse 1968). Because male warblers are more conspicuous and were therefore sampled more often, combining data for both genders may have overestimated mean foraging heights for the yellowrumped and Townsend's warblers. To explore species-specific movement patterns across riparian buffer widths, I combined treatments into three classes: Narrow (Very Narrow + Narrow buffers), Wide (Medium + Wide buffers), and Uncut Forest. I also reduced the eight movement angles to three classes: movements parallel to stream (1, 5), movements perpendicular to stream (3, 7), and movements diagonal to stream (2, 4, 6, 8) (Figure 2.1b). For each o f the four study species, I calculated the mean distance moved in each direction class in each treatment. I then used the Kruskal-Wallis non-parametric analysis o f variance to test for differences i n movement distance (1) between treatment classes and (2) within each treatment class (SPSS Inc. 1999). Because the analyses were exploratory, and the statistics were used as descriptors rather than for hypothesis testing, a significance level o f a = 0.10 was used for each test.  51  RESULTS Forest Structure General Description The total basal area/ha of live trees within the forest increased with riparian buffer width, reaching a maximum in the wide buffers and the unharvested forest (Figure 2.2a). The unharvested forest had significantly greater basal area than did the very narrow buffers (F =5.21, P<0.05). Canopy cover also increased with buffer 4>5  width to reach a maximum in the unharvested forest. The wide buffers and the unharvested forest had significantly greater canopy cover than the very narrow buffers (F , =10.37, P<0.01) (Figure 2.2b). 4 5  Primary tree species in the riparian buffers were hybrid white spruce and subalpine fir, with some lodgepole pine and Douglas-fir. As an indication of habitat diversity, the change in canopy composition with increasing distance from the stream was assessed in medium buffers, wide buffers and the unharvested forest. Although there was little change in composition with distance from the stream in medium buffers (Figure 2.3a), both wide buffers and the unharvested forest had less hybrid white spruce and more lodgepole pine and Douglas-fir (Figures 2.3b, 2.3c). Shrub cover increased with riparian buffer width to reach a maximum in the unharvested forest, though differences across treatments were not significant. Low shrubs (<1.0 m) reached their highest cover in the narrow buffers; conversely, shrubs >1.0 m had their greatest cover in the unharvested forest. The cover of coniferous shrubs increased steadily with buffer width; the cover of deciduous shrubs varied across treatments, but was greatest in the unharvested forest (Figure 2.4); Coniferous  52  shrubs included tree seedlings and saplings; common deciduous species included falsebox, black gooseberry (Ribes lacustre (Pers.) Poir.), black huckleberry and thimbleberry (Rubus parviflorus Nutt.). Although ground cover varied across treatments, the effect of riparian buffer width was significant only for coarse woody debris (Table 2.1). Plots in the unharvested forest contained the lowesjt mean cover of coarse woody debris and were significantly different (F =9.85, P<0.01) from the wide buffers, which possessed the 4i5  greatest mean cover. The unharvested forest also had the lowest mean cover of bare ground and litter, and the greatest mean cover of moss.  Width Effects on Forest Structure Twenty-three vegetation variables were used for factor analyses, though only 12 variables had component loadings greater ± 0.55 in the final analysis (Table 2.2). Four factors explained 68% of the variation in the vegetation data. Each variable loaded highly on one factor only, facilitating interpretation of each factor. Decreasing coarse woody debris cover and increasing basal area and canopy cover represented the first factor, which I described as  Tree Cover. The second factor, called  Coniferous Shrub Cover and Richness, was associated with increasing coniferous shrub cover, number of coniferous seedling/sapling species and number of shrub layers. The third factor was described as Herb  Cover and had a positive loading of  herbaceous ground cover and a negative loading of litter cover. The fourth factor,  called Deciduous Shrub Cover and Richness, had increasing deciduous shrub cover and species richness, increasing grass cover and decreasing moss cover.  53  Together, the first two factors (Tree Cover + Coniferous Shrub Cover/Richness) explained 44.6% of variation in the data and exhibited the greatest change across treatments. In general, wide buffers and the unharvested forest had higher scores along both axes than did narrow and very narrow buffers (Figure 2.5). Vegetation plots in the medium buffers were relatively evenly distributed along both axes. Regression analyses of factor scores against riparian buffer width yielded a positive and significant relationship between width and Tree Cover. Coniferous Shrub Cover/Richness also exhibited a positive, though not significant, relationship with width. There was no relationship between buffer width and either of Herb Cover or Deciduous Shrub Cover/Richness. The logarithmic model gave the best fit for Tree Cover versus buffer width (P<0.001), explaining 87% of the variation in the data (Figure 2.6). Both Tree Cover and Coniferous Shrub Cover/Richness increased rapidly with small increases in buffer width, and then slowed once a width of approximately 35 m was reached.  Songbird-Habitat Associations Songbird Density and Forest Structure I used partial correlations to determine the influence of each factor on the richness and density of forest songbird species within the study grids, while controlling for the influence of buffer width. Scatterplots indicated a positive linear relationship between forest songbird richness/density and each of Tree Cover and Coniferous Shrub Cover/Richness. Neither songbird variable showed a relationship with either of Herb Cover or Deciduous Shrub Cover/Richness. Once partial  54  correlations removed the effect o f buffer width, neither forest species richness nor density had a significant relationship with any o f the four factors (Table 2.3). Partial correlations were also used to test the width-independent influence o f forest structure on the density o f golden-crowned kinglets, Townsend's warblers and winter wrens. Scatterplots showed a positive linear relationship between the density of both golden-crowned kinglets and Townsend's warblers and each o f Tree Cover and Coniferous Shrub Cover/Richness. Winter wren density did not exhibit a distinct relationship with any o f the four factors. Once the influence o f buffer width was controlled for, the density o f Townsend's warblers and winter wrens was not significantly related to any o f the forest structure factors (Table 2.3). B y contrast, the density o f golden-crowned kinglets was positively and significantly related to Coniferous Shrub Cover and Richness.  Habitat Use and Movements A total o f 165 minutes o f foraging activity was recorded for the four species. Winter wrens were observed at 8 sites for a total of 21 minutes. Yellow-rumped warblers were recorded at 7 sites for 73 minutes. Golden-crowned kinglets were observed at 7 sites for a total o f 41 minutes, and Townsend's warblers were recorded at 4 sites for a total of 30 minutes. Foraging winter wrens gleaned from bark, ground and foliage (Figure 2.7a). The greatest proportion o f foraging time was spent on coarse woody debris and on the ground; root masses and mature Douglas-fir trees were used to a lesser extent (Figure 2.7b). Winter wrens were only recorded foraging in the uncut forest and the riparian buffers, though they were occasionally observed singing in the clearcuts.  55  Yellow-rumped warblers were frequently observed foliage gleaning and singing concurrently (Figure 2.8a). Foliage gleaning was occasionally interspersed with hover-gleaning; bark and ground gleaning were observed less often. Mature and pole-sized trees, mainly Douglas-fir, subalpine fir and lodgepole pine, were the most frequently used foraging substrates (Figure 2.8b). The mid-canopy layer was used for bark gleaning, and both mid and upper canopy layers were used for foliage gleaning and singing (Figure 2.8c). Yellow-rumped warblers foraged in all habitat types, though the primary habitats used were the riparian buffers, buffer edges and residual tree patches in clearcuts. Golden-crowned kinglets were most often observed foliage gleaning, though hover-gleaning was interspersed (Figure 2.9a). Mature and pole-sized trees, primarily hybrid white spruce and subalpine fir, were used extensively for foraging (Figure 2.9b). The mid-canopy layer was utilized for all types of gleaning; foraging was infrequently observed in the lower and upper canopy layers (Figure 2.9c). Goldencrowned kinglets spent most of their time foraging in the unharvested forest and riparian buffers, though they were also observed along buffer edges and in residual tree patches dispersed throughout the adjacent clearcut. Townsend's warblers were observed using a combination of foliage gleaning and hover-gleaning (Figure 2.10a). Foraging occurred on both pole-sized and mature hybrid white spruce and subalpine fir (Figure 2.10b); individuals used the lower, mid and upper canopy layers equally (Figure 2.10c). Townsend's warblers were occasionally observed foraging along buffer edges, but were most often seen in the riparian buffers and the unharvested forest.  56  Buffer width appeared to influence the movement patterns of winter wrens (Figure 2.1 la). Across treatments, wrens moved more in the perpendicular direction in the unharvested forest than in the narrow and wide buffers (^8.35, df = 2, P<0.05). Distances moved in the parallel and diagonal directions were similar across treatments. Within each treatment, distances moved in each direction were different in the narrow buffers (^=20.17, df = 2, PO.01) and the wide buffers (^=6.64, df = 2, P<0.05), but not in the unharvested forest. In both the narrow and wide buffers, wrens moved mostly in the upstream and downstream direction, rather than towards and away from the stream. Buffer width did not seem to strongly affect the movement patterns of yellowrumped warblers (Figure 2.1 lb). Across treatments, individuals moved more upstream and downstream in narrow buffers than in wide buffers, but the difference was not significant. Distances moved in both the perpendicular and the diagonal directions were similar across treatments. Within each treatment, the distances moved in each direction were significantly different in the narrow buffers (X=5.23, df = 2, P<0.10), but not in the wide buffers. Yellow-rumped warblers moved equally in all directions in wide buffers, but mostly upstream and downstream in narrow buffers Buffer width also did not seem to have a strong influence on the movement patterns of golden-crowned kinglets (Figure 2.12a). Between treatments, individuals moved greater distances upstream and downstream in the wide and narrow buffers than in the unharvested forest, but the difference was not significant. Distances moved perpendicular and diagonally to the stream were similar across treatments.  57  Within each treatment, the distances moved in each direction were not significantly different, though kinglets in both narrow and wide buffers moved the least in the perpendicular direction and the most in the upstream and downstream directions. Buffer width influenced the movement patterns o f Townsend's warblers (Figure 2.12b). Across treatments, individuals moved significantly more perpendicular to the stream in the unharvested forest than in the buffers (X=3.16, df = 1, P<0.10). Distances moved in the parallel and diagonal directions were similar across treatments. Within each treatment, the distances moved in each direction were significantly different in the buffers (X=4.83, df = 2, P O . 1 0 ) , but not in the unharvested forest. Within the buffers, Townsend's warblers moved mostly in the upstream and downstream direction, rather than towards and away from the stream.  58  DISCUSSION  Width Effects on Forest Structure Changes in forest vegetation occurred with increasing riparian buffer width, particularly in the canopy and understory layers. Wider buffers possessed higher moss, tall shrub and canopy cover, more vegetation layers and species, and larger trees than did narrower buffers. Because distance from the edge increased with increasing buffer width, these changes in vegetation were likely due to edge effects. Edge effects are often experienced along forest-clearcut boundaries (Voller 1998). In the Pacific Northwest, forest edges exhibit reduced canopy cover, stem density and basal area, and elevated rates of tree mortality and windthrow (Chen et al. 1992). In my study, windthrow along buffer edges contributed to lower basal area and canopy cover, and ground disturbance (from harvesting) and increased light levels likely resulted in the low moss and high litter cover observed in the narrower buffers. Given that edge effects on vegetation do not appear to penetrate as far into the forest as microclimate effects (Kremsater and Bunnell 1999), the narrower buffers in my study likely consisted entirely of edge habitat, whereas the wider buffers were more characteristic of interior forest habitat. Forest structure, as described by factor analysis, also exhibited the influence of edge effects. The first factor (Tree Cover) was significantly and positively related to buffer width; the second factor (Coniferous Shrub Cover and Richness) also had a positive, though not significant, relationship with buffer width. Tree Cover essentially described features of undisturbed forest habitat and separated narrow buffers from wide buffers and the unharvested forest. Coniferous Shrub Cover and  59  Richness broadly described the increased regeneration that has been shown to occur within the first 30 - 60 m of a forest-clearcut boundary (Chen et al. 1992). Overall, my results indicate that edge effects on forest vegetation decrease as riparian buffer width (and therefore distance from the edge) increases.  Songbirds and Buffer Width vs. Forest Structure The richness of forest songbird species was not significantly related to forest structure. Similar to the area effects noted by Freemark and Merriam (1986), Blake and Karr (1987) and Askins et al. (1987), buffer width was the primary influence on species richness. A positive relationship between species richness and habitat area has been observed by many researchers (Galli et al. 1976, Forman et al. 1976, reviewed in Connor and McCoy 1979, Wiens 1989). The density of forest songbird species was also not significantly related to forest structure. Similar results were described by Askins et al. (1987) and Doherty et al. (2000), who found that patterns of songbird density in forest fragments were strongly associated with fragment area and were unrelated to vegetation structure. Likewise, Ambuel and Temple (1983) were unable to find any area-related changes in vegetation structure that influenced songbird abundance. The densities of two common forest species, the winter wren and the Townsend's warbler, were highly correlated with riparian buffer width but not with forest structure. Given the strong relationship between density and buffer width for each species (Chapter 1), associations with habitat structure may have been too subtle to detect. However, other research into the habitats used by these species has indicated that such relationships do exist. In the central interior of B.C., Davis et al.  60  (1999) found Townsend's warblers to be positively correlated with the abundance of snags, and winter wrens to be positively correlated with total basal area, coarse woody debris volume and snag abundance. Hansen et al. (1995) showed that winter wren abundance was related to the total density of large diameter trees. Hagar et al. (1996) also found wren abundance to be correlated with the density of large diameter conifers and hardwoods, as well as the cover of coniferous shrubs. Although each model was developed from a different study site with a different set of measured variables, they all point to the importance of various canopy attributes. The density of golden-crowned kinglets was related both to buffer width and to a descriptor of coniferous shrub cover and richness. Davis et al. (1999) found that golden-crowned kinglets were associated with total basal area, basal area of large trees and snag abundance. Hansen et al. (1995) showed that golden-crowned kinglets were significantly correlated with total tree density, and Hagar et al. (1996) found that kinglets were related to the percent cover of pole-sized and mature trees. In light of these models, it is difficult to explain why the density of a canopy-dwelling species like the golden-crowned kinglet would be associated with the percent cover of coniferous shrubs (seedlings and saplings). It is possible that the increased cover of coniferous shrubs, mainly saplings <10 cm dbh, afforded more foraging opportunities for kinglets. It is also possible that the cover of coniferous shrubs was related to some other aspect of forest structure that was not measured and with which goldencrowned kinglets were associated. Although my analyses did not reveal distinct relationships between songbirds and forest structure, the influence of edge effects on buffer vegetation may in part  61  explain the pattern of increasing forest bird density and richness with buffer width. The long, thin medium and narrow buffers had a high ratio of edge to area. Because they possessed lower canopy cover, basal area and shrub cover than the wide buffers and the unharvested forest, they presented fewer foraging and nesting opportunities for some forest songbirds. In addition to the constraints of buffer width, edge effects and decreased resource availability may have further limited species presence and territory density in narrower buffers.  Habitat Use and Movements My observations of foraging songbirds indicated that certain habitat features were used differently by the four study species (winter wren, yellow-rumped warbler, Townsend's warbler, golden-crowned kinglet). By incorporating this data with habitat use information from other sources, species-specific inferences may be made regarding the potential effects of area-related vegetation changes in riparian buffers. Winter wrens prefer dark, relatively dense forests with abundant coarse woody debris, root masses and understory (Godfrey 1986, Davis et al.1995, Campbell et al. 1997). However, they are flexible with respect to habitat type and will also occur along edges, in logging slash and in more open forests with shrubby understory (Campbell et al. 1997). In my study area, winter wrens favoured coarse woody debris and root masses for foraging. Because winter wrens spend much of their time ground gleaning (Sabo and Holmes 1983) and nest in root masses, tree cavities and creek banks (Waterhouse 1998), habitat elements such as downed wood, shrubs and snags are important to them. The ability of this species to utilize a variety of habitat types may make it less sensitive to logging in and adjacent to riparian areas. At the same  62  time, the lower basal area, canopy cover and shrub cover found in riparian buffers may decrease the suitability of buffer habitat for wrens. The removal of snags from buffers (to meet safety regulations) can lead to the loss of potential nesting locations, though this loss may be offset by the increased amount of downed wood that results from windthrow. Overall, winter wrens may not be seriously affected by the changes in forest structure that occur in riparian buffers, as long as critical elements (i.e., canopy cover, wildlife trees, shrub cover) are retained. The yellow-rumped warbler is found in relatively open coniferous and mixedwood stands (Godfrey 1986) and exhibits little habitat specificity (Hunt and Flaspohler 1998). This species demonstrates foraging plasticity - the ability to 'make the best' of a bad situation by exploiting resources in different ways as forest conditions change (Morse 1968). The yellow-rumped warbler is a generalist arboreal forager that uses all parts of the tree (Hunt and Flaspohler 1998), though higher use of the mid and lower canopy layers has been documented (MacArthur 1958, Sabo and Holmes 1983). In my study area, yellow-rumped warblers were flexible in their choice of substrate, tree species, canopy strata and habitat type for foraging. They were observed most often on pole-sized and mature trees and used both mid and upper canopy layers. Because of their preference for more open conifer stands at high elevations (Campbell et al. 2001) and their ability to forage in a broad range of microhabitats, yellow-rumped warblers may be unaffected by the changes in canopy structure that occur in riparian buffers as long as some larger trees are retained (Hunt and Flaspohler 1998).  63  Townsend's warblers occupy fir and montane spruce-fir forests, favouring late successional unharvested stands (Wright et al. 1998). This species has been positively correlated with crown closure, tall conifers, high basal area, number of coniferous saplings and dense deciduous understory (Campbell et al. 2001, Wright et al. 1998). Townsend's warblers forage primarily in the upper canopy, on both coniferous and deciduous trees (Wright et al. 1998, Godfrey 1986). In my study area, they used all canopy layers and foraged both on pole and mature trees. Because of their preference for dense, mature forests with abundant understory, this species may be adversely affected by the structural changes that occur in riparian buffers. In addition, any selective harvesting of tall, large diameter conifers in buffers will remove preferred nesting and foraging locations (Wright et al. 1998). However, the fact that Townsend's warblers were recorded at low density even in narrow buffers (Chapter 1) suggests that individuals will occupy riparian buffers as long as important habitat features (i.e., moderate canopy cover, some larger trees) are maintained. Golden-crowned kinglets breed in a wide variety of forest types, but prefer boreal and subalpine forests (Ingold and Galati 1997). Although this species is usually found in dense, old growth or mature second growth forests (Campbell et al. 1997), it also occurs in logged areas, open and mixed-wood forests (Ingold and Galati 1997). Golden-crowned kinglets exploit a range of foraging maneuvers and substrates, focusing mostly on the lower and mid canopy (Sabo 1980), but also using the upper canopy layer (Sabo and Holmes 1983). In my study area, individuals foraged primarily in the mid canopy layer, using both pole-sized and mature trees. While golden-crowned kinglets are reliant on the canopy for foraging and nesting,  64  their flexibility with respect to forest type and foraging location suggests that they may not be particularly vulnerable to area-related vegetation changes in riparian buffers. The presence of golden-crowned kinglets in narrow buffers, albeit at low density (Chapter 1), indicates that they are tolerant of reduced canopy cover and basal area, provided that some mature trees are retained. The retention of trees and tree patches in cutblocks may also provide additional habitat for certain species. Both golden-crowned kinglets and yellowrumped warblers were observed foraging in retention patches within clearcuts. In particular, yellow-rumped warblers used both single trees and tree patches for foraging and singing perches. Residual trees can make cutblock habitat more attractive to forest bird species (Norton and Hannon 1997, Schieck et al. 2000, Schieck and Hobson 2000). In western boreal forests, harvest areas with tree retention patches (especially hardwoods) contain bird communities similar to those found in older forests (Schieck et al. 2000). Although larger patches attract more forest species than do smaller patches (Schieck and Hobson 2000), even the retention of single trees can provide habitat features for some forest species. Differences in songbird movement patterns were observed between riparian buffers and the unharvested forest. Overall, birds foraging in riparian buffers appeared to be constrained by buffer width. In unharvested stands (and wide buffers for yellow-rumped warblers), birds foraged in all directions equally; individuals moved about the same distance upstream and downstream as they did towards and away from the stream. In buffers, movements into the upland were truncated; individuals tended to move greater distances upstream and downstream than they did  65  towards and away from the stream. Robinson and Holmes (1982) found that foraging birds moved just far enough to take them into new search areas. Riparian buffers may therefore increase the energy demands of individuals that are forced to seek prey over longer distances than in undisturbed forests.  Conclusions The changes in vegetation that were observed across riparian buffers of different widths were likely the result of edge effects. Although changes in forest structure and forest songbird density were related to buffer width, they did not appear to be correlated with each other. Two possible reasons for this exist: (1) the appropriate vegetation variables were not measured, or were measured at too coarse a scale, or (2) forest birds were responding to edge effects besides those related to forest structure (e.g., competitive interaction with open-habitat species, increased predation, etc.) The foraging behaviours and habitat requirements of the species observed in my study and in others indicate a certain degree of flexibility. Most common forest species can probably adapt to the variations in vegetation structure that occur in riparian buffers, provided that critical habitat elements like large trees, moderate canopy cover and downed wood are retained. However, the needs of specialist species were not addressed by my study. These species are likely to be less flexible and therefore more vulnerable to changes in forest structure. Their general absence from riparian buffers in the study area may reflect an inability to adjust to habitat modifications.  66  cu 1. o  fc +u  s w c  0\  in  CN  co  O  cu ^ rH  f  r  •O  oo „ CN CN  ^  ^  ^  ro"  CO w  w  M  in  in CN  N  ^  in  H  N  • co  W  o  ss«  CN ^  O pa s- <  «  c  o  ^  , ^ oo >0  o  o  ^  CN s CN CO  ° £ co m co  *o  00  CD >.  ? O OO  to <u cu o S .-to U ffl J  'o  cx oo oo oo oo .2 . U  3 O o O g CD, CU & CO, O Moo So '  S  -5  CU  52  oo  «  2 O  ffi a S  CO  o  on in  m  o U s 1-  S  2  C  Q  -3  VO  vo  r-in  r-©  m  o  w  o s  % C/3  ^ Tf  ,  ,  ON  ON  cs o <=> • O 00 ^ ^ vo  CU  CO O CD  c o  m oo  ON OO  £  o o 43  CO  0  ON  vo ^ °^ J m h  co CD  •c > O  e«J  -*-» CD  60 CD >  "3  o ri o  <u ©  U s x:  —i CN 00 oo ©  u  ro »O g °^ 00 V  VO VO  «2|  ©  "S o U  '+->  O i-i X  B cd  ON  CT"  r-  VO  © ©  W J3  CO  t-- r-  CO  CN CN  &0  >,  a a  io +-» o  CJ  CD  c£-i o 3  co  a!  00  C  © .22 5 .2  1  CU La OX) es  00  <L)  >  .g  O O  •a  °  &  T3  CD CD  >~3  c  CD  a  > o  «  CJ  u  CJ  CJ  cu  o  ci)  ^ o  5  CO  cK  S  c o a 3 o o U % U  cn 3  O co T3  CD  o o  CD >  CD  o  > e  CJ  S co u a  CJ CD  X  oo  1  g"  1  CJ CJ  o >  NO ©N  o H  68  CO  5  *'  "3 to £ ,ca o ca ^  CU  fa  X5 +3  as ^ p © cu  1-  © U X!  P * . to  ON  to -d S <2 S  to fa o  -to c2 O  ft  b  •to  <U  O  cu ca (U  CN  o  0 CO  O  00 r-H  0  II  O  CO ^0 0  0  II  O  fa  II  O  fa  fa  CO  ^  00 co  CN  O  CO  Q  CU  cu o to •3 c .SP  s-  in o  CO  cu  IT)  -° fa o  •^f  o  ON  °. ©  m • o o  o o  in  :  f-i  II  0  fa  ' fa  •-'too co  CO CN  in  cu cu  £ 8 J  00  o r-o  co  5 to 03 O (J ' rt  co +5 — /  s  rto cu o ca — < + « v <u ^ A <-> °  6  '55  >  H  d  o . ft!  C C d ^  o  a  to 13 cu co  CU >  o U  X!  a  ON  CM CN CN  05  m in  CN  00  ©  CN  o  m  o  00  o o  m  1-  e o U  to  8 § "2  ^•s  9  ON  o ca -s  •n s « .S *3 u  co ID 3  CU  3 O  C  o  ^  CN  ©  fa  H  £  in o 0 II fa  ON  ON O  CO  ^  in o o II fa  in  O CO  n ?  O  O  CU  § b ca 13 .2 la 0  0  to  t:  o ca — cu ft ca .to cu TH  CO  ca * — »•  fa  XI  «  • ^ -2 n j -g X! 88  cS  >  <t| J3 8 3  cy2  Xi  .2  to cu  -a  co  ca  c  CO  >  •o CO  n o  to cu T3  CU  w  Xi  ju  CO  CU  '3  CU fa CO CO CU  O  fa  cu '3 CU  fa  co +J CO  CU 1-1  O  fa  "5b to  CO  cu o  T3  )-<  O  I  to CU  a  to cu to CO  to cu T3  o H  CO  o II  fa  > "£ Ji2 rH  in  00  to cu  T3  to cu cu to  stream  B.  Parallel (upstream/downstream)  Perpendicular (towards/away)  Diagonal (towards/away)  Figure 2.1. (a) Direction angles (relative to stream) used to record songbird movements, and (b) movement classes used for analysis.  S3  u  < c/l  Uncut  Wide  Medium  Narrow  V Narrow  Uncut  Wide  Medium  Narrow  V Narrow  B. 100%  80% o U  © s  60%  40%  A  20%  Buffer Type  Figure 2.2. (a) Mean basal area (± 1 SE), and (b) mean % canopy cover (± 1 SE) forest plots across treatments. Letters indicate significantly different treatments, based on one-way A N O V A with Bonferroni multiple comparisons.  Figure 2.3. Changes in % tree species composition with distance from stream in ( medium buffers, (b) wide buffers, and (c) the uncut forest. Sxw = hybrid white spruce; B l = subalpine fir; PI = lodgepole pine; F d = Douglas-fir.  50%  40%  >  ©  30%  s sx:  20%  6  s  10% -  Uncut  Wide  Medium  Coniferous  Narrow  V. Narrow  —o— Deciduous  Buffer Type  Figure 2.4. Mean % cover of coniferous and deciduous shrubs (± 1 SE) in forest plots across treatments.  73  • • • •  •  •  •  " •V i • v • A  •  v  •  1  "Hi A  O  Buffer Type  a  •  V  -1 - #  v  A v  O A *  * 1  V  v  A  * 1  Wide  v Medium A  i  Uncut  O Narrow i  •* Very Narrow  -2  Coniferous Shrub Cover & Richness (Increasing)  Figure 2.5. Scatterplot of Factor 1 (Tree Cover) scores against Factor 2 (Coniferous Shrub Cover + Richness) scores, by buffer type.  74  Buffer Width (m)  Figure 2.6. Relationship between riparian buffer width and Tree Cover (Factor 1 scores + 2).  75  60%  50%  40%  30%  G G = ground gleaning B G = bark gleaning F G = foliage gleaning CF = combination foraging  20%  "ITT  10%  K  0% G G  B G  F G  CF  Foraging Method  B. 60%  50%  40%  GR = ground CWD = coarse woody debris R M = root mass ST = sapling tree PT = pole tree M T = mature tree SN = snag  30%  20%  10%  0% GR  C W D  RM  S T  FT  M T  S N  Foraging Substrate  Figure 2.7. Winter wren activity budget. Proportion of time spent (a) engaged in foraging activities, and (b) foraging on different substrate types.  76  50%  40%  G G = ground gleaning B G = bark gleaning F G * = interspersed foliage + hover gleaning FG/S = concurrent foliage gleaning and singing  30%  20%  10%  0% G G  B G  FG/S  F G *  Foraging Method  50%  40%  GR = ground C W D = coarse woody debris SH = deciduous shrub ST = sapling tree PT = pole tree M T = mature tree SN = snag  30%  20%  10%  0% G R  S H  C W D  S T  PT  M T  S N  Foraging Substrate  80%  60%  40% MS = mid shrub L B = lower tree bole L C = lower canopy M C = mid canopy U C = upper canopy  20%  0%  MS  LB  LC  MC  uc  Vegetation Layers Figure 2.8. Yellow-rumped warbler activity budget. Proportion of time spent (a) engaged in foraging activities, (b) foraging on different substrate types, and (c) foraging in different vegetation layers.  77  100%  & '>  8 0 %  CJ  "5  6 0 %  a •G B  B G = bark gleaning H G = hover gleaning F G * = interspersed foliage + hover gleaning  4 0 %  a>  a  20%  0% BG  HG  FG*  Foraging Method B. 60%  03 S-  s  i§iii§ 40%  CO I*  <u a •*•>  B  20%  a>  a co  1  CU  E  Pllll  M T = mature tree PT = pole tree ST = sapling tree  0% ST  FT  M T  Foraging Substrate  60%  -  cu a  40%  B  CU  a. co  * fa -  20%  CU  L C = lower canopy M C = mid canopy U C = upper canopy  0% LC  M C  U C  Vegetation Layers Figure 2.9. Golden-crowned kinglet activity budget. Proportion of time spent (a) engaged in foraging activities, (b) foraging on different substrate types, and (c) foraging in different vegetation layers.  78  A. 100%  +^  '>  < CU  a a cu CU  t/3  80%  60%  40%  H G = hover-gleaning F G * = interspersed foliage + hover  20%  <U  s  0% HG  F G  Foraging Method B. 00%  s. CA  s I.  cu a cu cu cu E  80%  60%  40%  SH = deciduous shrub PT = pole tree M T = mature tree  20%  H 0%  PT  SH  MT  Foraging Substrate C. 50%  cu « J u cu a e cu Q. c/3  cu B H  40%  i  30%  20%  US = upper shrub L C = lower canopy M C = mid canopy U C = upper canopy  10%  0%  us  LC  M C  uc  Vegetation Layers Figure 2.10. Townsend's warbler activity budget. Proportion of time spent (a) engaged in foraging activities, (b) foraging on different substrate types, and (c) foraging in different vegetation layers.  79  A. 25  20  15  10  I  Parallel  Perpendicular  Diagonal  o  B.  • Narrow (n=11) • Wide (n=5) • Uncut (n=5)  t/3  Parallel  Perpendicular •  Narrow (n=7)  •  Diagonal  Wide (n=7)  Direction Moved Figure 2.11. Mean distance moved (± 1 SE) in directions relative to stream by (a) winter wrens, and (b) yellow-rumped warblers. Sample sizes are given in each legend. (Parallel = upstream/downstream; Perpendicular = directly towards/away from stream; Diagonal = towards/away from stream at an angle).  80  A. 20  15  10  J  >  Parallel  Perpendicular  Diagonal  ©  • Narrow (n=4) CU CU  =  • Wide (n=3)  • Uncut (n=4)  B.  E/3  Parallel  Perpendicular • Buffer (n=3)  Diagonal  • Uncut (n=2)  Direction Moved Figure 2.12. Mean distance moved (± 1 SE) in directions relative to stream by (a) golden-crowned kinglets, and (b) Townsend's warblers. Sample sizes are given in each legend. (Parallel = upstream/downstream; Perpendicular = directly towards/away from stream; Diagonal = towards/away from stream at an angle).  81  CONCLUSIONS A N D RECOMMENDATIONS Species are frequently classified based on their response to landscape structure or habitat fragmentation, but the data to support these classifications are often lacking (Villard 1998). For example, empirical evidence for edge avoidance in species thought to depend on forest interior is scant and few species appear to be "true" forest-interior specialists (Villard 1998). Most forest songbirds in my study area did not demonstrate the same degree of edge avoidance and area-sensitivity as found in other studies (e.g., Darveau et al. 1995, Gyug 1996, Hagar 1999, Whitaker and Montevecchi 1999, Pearson and Manuwal). Changes in the density of different forest species basically tracked changes in buffer width; there were no obvious thresholds below which the loss of forest species was disproportionately greater than the loss of buffer habitat. There are several possible landscape-level explanations for these stand-level observations. First, the study area is still surrounded by forest at both the local and the regional level. Proximity to other local populations can influence the occupancy of forest fragments by birds (Villard 1998). Both Trzcinski et al. (1999) and McGarigal and McComb (1995) found that the distribution of forest songbirds was positively correlated with forest cover across the landscape. Askins et al. (1987) found that the density of "forest-interior" birds was higher in heavily forested regions, concluding that the dispersal of birds from surrounding forests is important to maintain populations in forest fragments. Second, the study area occurs in a naturally heterogeneous landscape that has been historically disturbed by fire and insect outbreaks (Province of B.C. 1995a). Because the forest birds inhabiting this  82  landscape have adapted to natural disturbance at various scales, they are likely more resilient to anthropogenic disturbances such as harvesting (Schieck et al. 1995, Schmiegelow et al. 1997). Third, many songbird declines in fragmented landscapes have been attributed to brown-headed cowbird parasitism (Brittingham and Temple 1983, May and Robinson 1985, Robinson 1992). There were no cowbirds recorded either in or around my study grids, leading me to conclude that brown-headed cowbirds were largely absent from the study area. Although confirmation of this would require a more detailed study with experienced observers, it is unlikely that cowbird nest parasitism is having a strong effect on the local songbird population. Overall, riparian buffers provided usable habitat for forest songbirds breeding in the study area. Buffers were used by the winter wren, a riparian-associated species, and by various non-riparian species. Even those species considered "upland specialists", such as the golden-crowned kinglet, Hammond's flycatcher and Townsend's warbler (McGarigal and McComb 1992, Martin and Wiebe 1997, Pearson and Manuwal 2001), established territories in the buffers. Riparian buffers > 60 m in width were used by almost every species observed in the unharvested forest, but even the narrower buffers provided some habitat for several forest species. Although wide riparian buffers are not likely to meet the needs of all forest songbirds, the range of buffer widths created by current forest practices appears to provide suitable habitat for many songbirds in the study area.  Study Limitations The restricted temporal scope of this study is one major limitation to the generality of my research. Songbird data were collected for one breeding season only  83  due to logistical constraints. North American songbird populations frequently exhibit moderate to large-scale annual fluctuations in species abundances (Folkard and Smith 1995, Schmiegelow et al. 1997) and in observed relationships between species abundance and habitat area (Blake and Karr 1987). The winter that preceded the sampling season was particularly long and harsh, likely resulting in decreased songbird numbers and possibly in the general absence of certain species that should have been more common (e.g., red-breasted nuthatch, brown creeper, hermit thrush). Thus, the extrapolation of conclusions from this one year of abnormal weather to other milder years is unreliable. The study also did not incorporate both pre- and post-treatment sampling; the absence of any pre-treatment data precludes the assumption of true replication. In addition, site tenacity in songbirds can produce local densities that reflect past, rather than present habitat quality (Van Home 1983). Without longer-term data for the study area, it is impossible to know whether or not the patterns exhibited during the 1999 breeding season are representative of general trends and true conditions. A second limitation of this study is the lack of any nesting or productivity data. The density of individual species and of all species combined was used to indicate the ability of riparian buffers to provide suitable breeding habitat. This assumes that bird density and habitat quality are positively correlated. Without any supporting nesting success and productivity data, however, this assumption may be misleading or even incorrect (Van Home 1983). The singing males recorded in the buffers may have been "floaters", subdominant males forced into sub-optimal habitat where they were unable to attract mates. Alternatively, males in the buffers may have  84  successfully attracted females, but the negative influence of edge effects subsequently resulted in low nesting or fledging success. Thus, riparian buffers thought to be of adequate width to maintain breeding songbirds may actually be acting as habitat sinks (Van Home 1983). Further research that incorporates measures of fitness would be required to determine whether or not riparian buffers truly contribute to viable populations of forest songbirds. A third limitation of the research is the transferability of the data and results. The relationships between species density and buffer width that were defined through regression analysis should be viewed as descriptive, rather than predictive. Because of the small sample size and low level of replication, there are not sufficient data to extrapolate these relationships beyond the study area or buffer widths sampled. In addition, the study examined streams with a riparian buffer on one side and contiguous forest on the other side. Although harvesting on one side of the stream is a common practice within T F L #35, it is not the method applied in all managed forests. If the buffers had been isolated by upland harvesting on both sides of the stream, the density of forest birds would have likely been much lower and some species (e.g., varied thrush) may have been entirely absent (but see Waterhouse and Harestad 1999). Patches that are too small to support certain species can do so only if additional, suitable habitat exists nearby (Blake and Karr 1987). Thus, the presence of unharvested forest across the stream allowed buffers to support more individuals and more species than would have occurred in isolated buffers. The conclusion that a minimum riparian buffer width of 30 m is needed to provide habitat for most forest  85  songbirds would not be valid if extrapolated to similar forest types with isolated riparian buffers.  Management Recommendations The recommendations listed below are intended to act as a guide for forest managers, particularly those working within T F L 35 and other, similar high-elevation forest types. Although these practices may not necessarily be employed across the entire managed landscape, they can be applied to portions of the landscape to meet specific management objectives. 1.  Variable-width buffer retention. Results of my research indicate that a minimum riparian buffer width of 30 m is required to provide habitat for many forest songbird species in the study area. In other studies however, the same forest species exhibited high variability with respect to width-sensitivity (Darveau et al 1995, Gyug 1996, Kinley and Newhouse 1997, Hagar 1999, Manuwal and Pearson 2001). This variability makes it difficult to propose an absolute minimum riparian buffer width. Instead, I recommend variablewidth buffers that make use of site-specific topography. For example, wide buffers can be assigned to riparian areas that are inoperable due to steep slopes in exchange for greater volumes extracted from riparian areas that may be particularly susceptible to windthrow. In this way, patches of wide forested buffers will be interspersed with sections of stream having little or no buffering - an alternative to having long, narrow buffers that provide little habitat and can be subject to high rates of blowdown.  86  2. Single-side harvesting. The contiguous forest on the opposite of each stream likely contributed to the songbird densities that were observed in the riparian buffers. Removing timber along only one side of a stream in each pass will increase the value of the retained buffer as usable habitat for forest songbirds. The isolation of buffers by concurrent harvesting on both sides of a stream is not recommended. 3. Retention of structure in clearcuts. Residual patches of live trees retained in clearcuts were utilized for foraging and singing by various songbird species, including dark-eyed juncos, yellow-rumped warblers, golden-crowned kinglets and Townsend's warblers. Single live trees and snags were also utilized for foraging and perching. The retention of living and dead deciduous and coniferous trees in harvested areas can provide important habitat elements to both open-habitat and forest songbird species. The continued maintenance of these structural features, in both dispersed and clumped distributions, is recommended. 4. Additional conservation strategies. Riparian buffers retained in harvested areas can mitigate the effects of habitat loss for numerous forest songbird species. The juxtaposition of clearcuts and forested buffers can even increase songbird richness and diversity across the landscape. However, the maximization of species richness and diversity should not be the only goal of management. It is necessary to consider the composition of the songbird community and the continued maintenance of both common and rare species. Even wide buffers will not support all species - songbirds with a high degree  87  of area-sensitivity will require large patches of unharvested forest for breeding habitat. Thus, it is important to consider the use of additional conservation strategies, such as Old Growth Management Areas (Ministry of Forests 1995), to sustain a complete forest bird community. Monitoring songbird numbers. Monitoring is an integral component of any wildlife management strategy. Recommendations made from this study are based on a single year of data and may not be sufficient to maintain songbird species and numbers over time. The implementation of a monitoring program, even a basic survey of songbird presence/absence every 2 or 3 years, is highly recommended. One cost-effective way to gather monitoring data is to request that a North American Breeding Bird Survey route be established within the management area. This would allow the collection of annual data, by an experienced observer, for the long-term monitoring of local population trends.  88  LITERATURE CITED Able, K . P . and B . R . Noon. 1976. A v i a n community structure along elevational gradients in the northeastern United States. Oecologia 26: 275 - 294. Altman, J. 1974. Observational study o f behaviour: Sampling methods. Behaviour 49: 2 2 7 - 2 6 7 . Ambuel, B . , and S.A. Temple. 1983. Area-dependent changes in the bird communities and vegetation o f southern Wisconsin forests. Ecology 64: 1057 - 1068. Anderson, S.H. and H . H . Shugart, Jr. 1974. Habitat selection o f breeding birds in an east Tennessee deciduous forest. Ecology '55: 828 - 837. Askins, R . A . , Philbrick, M . J . , and D.S. Sugeno. 1987. Relationships between the regional abundance of forest and the composition of forest bird communities. Biological Conservation 39: 129 - 152. Askins, R . A . , Lynch, J.F., and R. Greenberg. 1990. Population declines in migratory birds in eastern North America. Current Ornithology 7 : 1 - 57. Austen, J.J.W., Francis, C M , Burke, D . M . and M . S . W . Bradstreet. 2001. Landscape context and fragmentation effects on forest birds in southern Ontario. The Condor 103: 701 - 7 1 4 . Beese, W . J . and A . A . Bryant. 1999. Effect of alternative silvicultural systems on vegetation and bird communities in coastal montane forests of British Columbia, Canada. Forest Ecology and Management 115: 231 - 242. Bibby, C.J., Burgess, N . D . and D . A . H i l l . 1992. B i r d census techniques. Academic Press: Harcourt Brace & Company Publishers, London, U K . Blake, J.G., and J.R. Karr. 1984. Species composition o f bird communities and the conservation benefit of large versus small forests. Biological Conservation 30: 173 187. Blake, J.G., and J.R. Karr. 1987. Breeding birds o f isolated woodlots: area and habitat relationships. Ecology 68: 1724 - 1734. Bonham, C D . 1989. Measurements for terrestrial vegetation. John Wiley & Sons. N e w York, N . Y . , U S A . Brand, L . A . and T . L . George. 2000. Predation risk o f nesting birds in coast redwood forest fragments. Journal o f Wildlife Management 64: 42 - 51.  89  Brand, L.A. and T.L. George. 2001. Response of passerine birds to forest edge in coast redwood forest fragments. The Auk 118: 678 - 686. Brittingham, M.C. and S.A.Temple. 1983. Have cowbirds caused forest songbirds to decline? Bioscience 33: 31 - 35. Brotons, L., Magrans, M . , Ferrus, L. and J. Nadal. 1998. Direct and indirect effects of pollution on the foraging behaviour of forest passerines during the breeding season. Canadian Journal of Zoology 76: 556 - 565. Brower, L . E . , Zar, J.M. and C.N. von Ende. 1990. Field and laboratory methods for general ecology. Third Edition. W. C. Brown Publishers, Dubuque, IA, USA. Bunnell, F. L . 1995. Forest-dwelling vertebrate faunas and natural fire regimes in British Columbia: patterns and implications for conservation. Conservation Biology 9: 636 - 644. Burke, D.M. and E. Nol. 2000. Landscape and fragment size effects on reproductive success of forest-breeding birds in Ontario. Ecological Applications 10: 1749 - 1761. Burnham, K.P. and D.R. Anderson. 1998. Model selection and inference: A practical information-theoretic approach. Springer-Verlag, New York, N Y , USA. Campbell, R.W., Dawe, N.K., McTaggart-Cown, I., Cooper, J.M., Kaiser, G.W., McNall, M.C.E. and G.E.J. Smith. 1997. The Birds of British Columbia, Volume III. Passerines: Flycatchers through Vireos. U B C Press, Vancouver, BC, Canada. Campbell, R.W., Dawe, N.K., McTaggart-Cown, II, Cooper, J.M., Kaiser, G.W., Stewart, A.C. and M.C.E. McNall. 2001. The Birds of British Columbia, Volume IV. Passerines: Wood Warblers through Old World Sparrows. U B C Press, Vancouver, BC, Canada. Campi, M.J. and R. Mac Nally. 2001. Birds on edge: avian assemblages along forestagricultural boundaries of central Victoria, Australia. Animal Conservation 4: 121 132. Chen, J., Franklin, J.F. and T.A. Spies. 1992. Vegetation responses to edge environments in old-growth Douglas-fir forests. Ecological Applications 2: 387 — 396. Chen, J., Franklin, J.F. and T.A. Spies. 1995. Growing-season microclimatic gradients from clearcut edges into old-growth Douglas-fir forests. Ecological Applications 5: 74-86. Clark, K., Euler, D. and E. Armstrong. 1983. Habitat associations of breeding birds in cottage and natural areas of central Ontario. Wilson Bulletin 95: 77 - 96.  90  Comrey, A . L . and H.B. Lee. 1992. A first course in factor analysis. Second Edition. Erlbaum, Hillsdale, NJ. Conner, E.F. and E.D. McCoy. 1979. The statistics and biology of the species-area relationship. The American Naturalist 113: 791 - 833. Cook, W.C. and J. Stubbendieck. 1986. Range research: Basic problems and techniques. Society for Range Management, USA. Croonquist, M.J. and R.P. Brooks. 1993. Effects of habitat disturbance on bird communities in riparian corridors. Soil and Water Conservation 48: 65-70. Darveau, M . , Beauchesne, P., Belanger, L., Hout, J. and P. Larue. 1995. Riparian forest strips as habitat for breeding birds in boreal forest. Journal of Wildlife Management 59: 67 - 78. Davidson, A.S. and R.L. Knight. 2001. Avian nest success and community composition in a western riparian forest. Journal of Wildlife Management 65: 334 344. Davis, L., Waterhouse, M . and H.Armleder. 1999. A comparison of the breeding bird communities in serai stages of the Engelmann Spruce-Subalpine fir zone in east central British Columbia. Res. Br., B.C. Ministry of Forests, Victoria, B.C. Working Paper #39/1999. Dickson, J.G., Williamson, J.H., Conner R.N. and B. Ortego. 1995. Streamside zones and breeding birds in eastern Texas. Wildlife Society Bulletin 23: 750 - 755. Doherty Jr, P.F. and T.C. Grubb, Jr. 2000. Habitat and landscape correlates of presence, density and species richness of birds wintering in forest fragments in Ohio. Wilson Bulletin 112: 388 - 394. Donovan, T . M . , Jones, P.W., Annand, E . M . and F.R. Thompson III. 1997. Variation in local-scale edge effects: mechanisms and landscape context. Ecology 78: 2064 2075. Ehrlich, P.R., Dobkin, D.S. and D. Wheye. 1988. The Birder's Handbook. Simon & Schuster Inc. New York, N.Y., USA. ESRI Inc. 1998. ArcView, Version 3.1. Faaborg, J. and W.J. Arendt. 1992. Long-term declines of winter resident warblers in a Puerto Rican dry forest: which species are in trouble? Pp. 57 - 74 in Ecology and conservation of Neotropical migrant landbirds. Hagan III, J.M. and D.W. Johnston (eds.). Smithsonian Institution Press, Washington, D.C.  91  Faaborg, J., Brittingham, M . , Donovan, T. and J. Blake. 1995. Habitat fragmentation in the temperate zone. In Ecology and management of Neotropical migratory birds. Martin, T.E. and D . M . Finch (eds.). Oxford University Press, New York, N Y , USA. Folkard, N.F.G. and J.N.M. Smith. 1995. Evidence for bottom-up effects in the boreal forest: Do passerine birds respond to large-scale experimental fertilization? Canadian Journal of Zoology 73: 2231 - 2237. Forman, R.T.T., Galli, A . E . and C F . Leek. 1976. Forest size and avian diversity in New Jersey woodlots with some land use implications. Oecologia (Berl.) 26: 1-8. Freemark, K. and B. Collins. 1992. Landscape ecology of birds breeding in temperate forest fragments. In Ecology and conservation of Neotropical migrant landbirds. Hagan III, J.M. and D.W. Johnston (eds.). Smithsonian Institution Press, Washington, USA. Freemark, K . E . and H.G. Merriam. 1986. Importance of area and habitat heterogeneity to bird assemblages in temperate forest fragments. Biological Conservation 36: 115 - 141. Galli, A . E . , Leek, C.F., and R.T. Forman. 1976. Avian distribution in forest islands of different sizes in central New Jersey. The Auk 93: 356 - 364. Gates, J.E. and L.W. Gysel. 1978. Avian nest dispersion and fledging success in fieldforest ecotones. Ecology 59: 871 - 883. Godfrey, W.E. 1986. The birds of Canada (Revised Edition). National Museums of Canada. Ottawa, Canada. Gyug, L.W. 1996. Timber harvesting effects on riparian wildlife and vegetation in the Okanagan Highlands of British Columbia. Part II: Breeding birds. B.C. Environment, Penticton, B.C. Hagar, J.C. 1999. Influence of riparian buffer width on bird assemblages in western Oregon. Journal of Wildlife Management 63: 484 - 496. Hagar, J . C , McComb, W . C and W.H.Emmingham. 1996. Bird communities in commercially thinned and unthinned Douglas-fir stands of western Oregon. Wildlife Society Bulletin 24: 252 - 266. Hansen, A.J., McComb, W . C , Vega, R., Raphael, M . G . and M . Hunter. 1995. Bird habitat relationships in natural and managed forests in the West Cascades of Oregon. Ecological Applications 5: 555 - 569.  92  Hansson, L. 1983. Bird numbers across edges between mature conifer forests and clearcuts in central Sweden. Ornis Scandanavia 14: 92 - 103. Harris, L.D. 1984. The fragmented forest: island biogeographic theory and the preservation of biotic diversity. University Chicago Press, Chicago, Illinois. Harris, L.D. 1988. Edge effects and conservation of biotic diversity. Conservation Biology 2: 330-332. Haskell, D.G. 1995. A reevaluation of the effects of forest fragmentation on rates of bird-nest parasitism. Conservation Biology 9: 1316-1318. Hicks, C.R. 1992. Fundamental Concepts in the Design of Experiments. Saunders College Publishing, New York, N.Y., USA. Hilborn, R. and M . Mangel. 1997. The ecological detective: Confronting models with data. Princeton University Press, Princeton, New Jersey, USA. Hobson, K . A . and E. Bayne. 2000. Effects of forest fragmentation by agriculture on avian communities in the southern boreal mixedwoods of western Canada. Wilson Bulletin 112: 373 -387. Hunt, P.D. and D.J. Flaspohler. 1998. Yellow-rumped warbler (Dendroica coronata). In The Bird of North America, No. 376. A. Poole and F. Gill (eds.). The Birds of North America, Inc., Philadelphia, PA. Hunter, M.L., Jr. 1990. Wildlife, forests and forestry: Principles of managing forests for biological diversity. Prentice-Hall, Englewood Cliffs, New Jersey, USA. Ingold, J.L. and R. Galati. 1997. Golden-crowned kinglet (Regulus satrapd). In The Birds of North America, No. 301. A. Poole and F. Gill (eds.). The Birds of North America, Inc., Philadelphia, PA. James, F.C. and N.O. Warner. 1982. Relationships between temperate forest bird communities and vegetation structure. Ecology 63: 159- 171. Johnson, W.N. and P.W. Brown. 1990. Avian use of a lakeshore buffer strip and an undisturbed lakeshore in Maine. Northern Journal of Applied Forestry 7: 114-117. Keller, C.M.E., Robbins, C.S. and J.S. Hatfield. 1993. Avian communities in riparian forests of different widths in Maryland and Delaware. Wetlands 13: 137 - 144. Kinley, T.A. and N..J. Newhouse. 1997. Relationship of riparian reserve zone width to bird density and diversity in southeastern British Columbia. Northwest Science 71: 75 - 86.  93  Krebs, C J . 1989. Ecological methodology. Harper Collins Publishers, New York, NY, USA. Kremsater, L. and F.L. Bunnell. 1999. Edge effects: Theory, evidence and implications to management of western North American forests. In Forest Wildlife and Fragmentation Management Implications. Rochelle, J.A., Lehmann, L.A. and J. Wisniewski (eds.). Koninklijke Brill NV, Leiden, The Netherlands. Kroodsma, R.L. 1982. Edge effect on breeding forest birds along a power-line corridor. Journal of Applied Ecology 19: 361 - 370. Lance, A.N. and M . Phinney. 2001. Bird responses to partial retention timber harvesting in central interior British Columbia. Forest Ecology and Management 142: 267-280. Lehner, P.N. 1987. Design and execution of animal behaviour research: An overview. Journal of Animal Science 65:1213 - 1219. Lloyd, D., Angove, K., Hope, G. and C. Thompson. 1990. A guide to site identification and interpretation for the Kamloops Forest Region. B.C. Ministry of Forests, Victoria, B.C. Lynch, J.F. and D.F. Whigham. 1984. Effects of forest fragmentation on breeding bird communities in Maryland, USA. Biological Conservation 28: 287 - 324. MacArthur, R.H. 1958. Population ecology of some warblers of northeastern coniferous forests. Ecology 39: 599 - 619. Mahoney, N., Nol, E. and T. Hutchinson. 1997. Food-chain chemistry, reproductive success, and foraging behaviour of songbirds in acidified maple forests of central Ontario. Canadian Journal of Zoology 75: 509 - 517. Manuwal, D.A. and A.B. Carey. 1991. Methods for measuring populations of small, diurnal forest birds. Gen. Tech. Rep. PNW-GTR-278. US Department of Agriculture, Forest Service, Pacific Northwest Research Station, Oregon, USA. Martin, T.E., Paine, C.R., Conway, W., Hochachka, M . , Allen, P. and W. Jenkins. 1997. BBIRD Field Protocol. Montana Cooperative Wildlife Research Unit. University of Montana, Missoula, Montana, USA. Maurer, B.A. and R.C. Whitmore. 1981. Foraging of five bird species in two forests with different vegetation structure. Wilson Bulletin 93: 478 - 490. May, R.M. and S.K. Robinson. 1985. Population dynamics of avian brood parasitism. American Naturalist 126: 475-494.  94  McGarigal, K. and W.C. McComb. 1992. Streamside versus upslope breeding bird communities in the central Oregon coast range. Journal of Wildlife Management 56: 10-20. McGarigal, K. and W.C. McComb. 1995. Relationships between landscape structure and breeding birds in the Oregon coast range. Ecological Monographs 65: 235 - 260. Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. Research Branch, B.C. Ministry of Forests, Victoria, Canada. Meiklejohn, B.A. and J.W. Hughes. 1999. Bird communities in riparian buffer strips of industrial forests. American Midland Naturalist 141: 172 - 184. Morrison, M.L., Timossi, C. and K.A. With. 1987. Development and testing of linear regression models predicting bird-habitat relationships. Journal of Wildlife Management 51: 247 - 253. Morse, D.H. 1968. American warblers: An ecological and behavioural perspective. Harvard University Press, Cambridge, Massachusetts, USA. Noon, B.R. 1980. Techniques for sampling avian habitats. Gen. Tech. Rep. GTR-87. U.S. Department of Agriculture, Forest Service, Rocky Mountain Experiment Station. Norton, M.R. and S.J. Hannon. 1997. Songbird response to partial-cut logging in the boreal mixedwood forest of Alberta. Canadian Journal of Forest Research 27: 44 53. Osborne, W.S. and K. Green. 1992. Seasonal changes in composition, abundance and foraging behaviour of birds in the Snowy Mountains. Emu 92: 93 - 105. Pearson, S.F. and D.A. Manuwal. 2001. Breeding bird response to riparian buffer width in managed Pacific Northwest Douglas-fir forests. Ecological Applications 11: 840-853. Province of British Columbia. 1995a. Forest Practices Code: Biodiversity Guidebook. B.C. Ministry of Forests and B.C. Environment, Victoria, Canada. Province of British Columbia. 1995b. Forest Practices Code: Riparian Management Guidelines. B.C. Ministry of Forests, Victoria, Canada. Robbins, C.S., Dawson, D.K., and B.A. Dowell. 1989. Habitat area requirements of breeding birds of the Middle Atlantic States. Wildlife Monologues 103: 1 -34. Robinson, S.K. and R.T. Holmes. 1982. Foraging behaviour of forest birds: the relationships among search tactics, diet, and habitat structure. Ecology 63: 1918 — 1931.  95  Robinson, S.K. 1992. Population dynamics of breeding birds in a fragmented Illinois landscape. Pp. 408 — 418 m Ecology and conservation of Neotropical migrant landbirds. Hagan III, J.M. and D.W. Johnston (eds.). Smithsonian Institution Press, Washington, D.C. Robinson, S.K., Rothstein, S.L, Brittingham, M . C , Petit, L.J. and J.A. Grzybowski. 1995. Ecology and behaviour of cowbirds and their impact on host populations. In Ecology and Management of Neotropical migratory birds. Martin, T.E. and D . M . Finch (eds.). Oxford University Press, New York, USA. Sabo, S.R. 1980. Niche and habitat relations in subalpine bird communities of the White Mountains of New Hampshire. Ecological Monographs 50: 241 - 259. Sabo, S.R. and R.T. Holmes. 1983. foraging niches and the structure of forest bird communities in contrasting montane habitats. Condor 85: 121 - 138. Sauer, J.R. and S. Droege. 1989. Geographic patterns in population trends of Neotropical migrants in North America. Pp. 26 - 42 in Ecology and conservation of Neotropical migrant landbirds. Hagan III, J.M. and D.W. Johnston (eds.). Smithsonian Institution Press, Washington, D.C. Schieck, J. and K. Hobson. 2000. Bird communities associated with live residual tree patches within cut blocks and burned habitat in mixedwood boreal forests. Canadian Journal of Forest Research 30: 1281 - 1295. Schieck, J., Lertzman, K., Nyberg, B., and R. Page. 1995. Effects of patch size on birds in old-growth montane forests. Conservation Biology 9: 1072 - 1084. Schieck, J., Stuart-Smith, K. and M . Norton. 2000. Bird communities are affected by amount and dispersion of vegetation retained in mixedwood boreal forest harvest areas. Forest Ecology and Management 126: 239 - 254. Schmiegelow, F.K.A., Machtans, C.S. and S.J. Hannon. 1997. Are boreal birds resilient to forest fragmentation? An experimental study of short-term community responses. Ecology 78: 1914 - 1932. Simberloff, D. 1994. Habitat fragmentation and population extinction of birds. Ibis 137: S 1 0 5 - S i l l . SigmaPlot SPSS Inc. 2001. SigmaPlot 2001, Version 7.0. Smith, R.J. and J.M. Schaefer. 1992. Avian characteristics of an urban riparian strip corridor. Wilson Bulletin 104: 732 - 738. SPSS Inc. 1998. SPSS Base 8.0 Applications Guide.  96  SPSS Inc. 1999. SPSS for Windows, Version 10.0.0. Stauffer, D.F. and L.B. Best. 1980. Habitat selection by birds of riparian communities: evaluating effects of habitat alterations. Journal of Wildlife Management 44: 1-15. Svensson, S. and S M . Taylor. 1970. Recommendations for an international standard for a mapping method in bird census work. Audubon Field Notes 24: 723 - 726. Swift, B.L., Larson, J.S. and R.M. DeGraaf. 1984. Relationship of breeding bird density and diversity to habitat variables in forested wetlands. Wilson Bulletin 96: 48 -59. Tabachnick, B.G. and L.S. Fidell. 1996. Using multivariate statistics. Third Edition Harper Collins College Publishers, New York, USA. Terborgh, J. 1989. Perspectives on the conservation of Neotropical migrant landbirds. Pp. 7 - 19 in Ecology and conservation of Neotropical migrant landbirds. Hagan III, J.M. and D.W. Johnston (eds.). Smithsonian Institution Press, Washington, D.C. Thurmond, D.P., Miller, K.V. and T.G. Harris. 1995. Effect of streamside management zone width on avifauna communities. Southern Journal of Applied Forestry 19: 166- 169. Trzcinski, M.K., L. Fahrig and G. Merriam. 1999. Independent effects of forest cover and fragmentation on the distribution of forest breeding birds. Ecological Applications 9: 586-593. Triquet, A . M . , McPeek, G.A. and W.C. McComb. 1990. Songbird diversity in clearcuts with and without a riparian buffer strip. Journal of Soil and Water Conservation 45: 500-503. Vander Haegen, W.M. and R.M. Degraaf. 1996. Predation on artificial nests in forested riparian buffer strips. Journal of Wildlife Management 60: 542 - 550. Van Home, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47: 893 - 901. Villard, M . 1998. On forest-interior species, edge avoidance, area sensitivity, and dogmas in avian conservation. The Auk 115: 801 - 805. Villard, M . , Trzcinski, M.K. and G. Merriam. 1999. Fragmentation effects on forest birds: Relative influence of woodland cover and configuration on landscape occupancy. Conservation Biology 12: 774 - 783.  97  Voller, J. 1998. Managing for edge effects. In Conservation biology principles for forested landscapes. Voller, J. and S. Harrison (eds.). U B C Press, Vancouver, Canada. Waterhouse, F.L. 1998. Habitat of winter wrens in riparian and upland areas of coastal forests. M.Sc. Thesis, Simon Fraser University, British Columbia, Canada. Waterhouse, F.L. and A.S. Harestad. 1999. The value of riparian forest buffers as habitat for birds in the Coastal Western Hemlock Zone, British Columbia: a pilot study. Forest Research Technical Report TR-001. B.C. Forest Service, Nanaimo, B.C. Whitaker, D.M., Carroll, A . L . and W.A. Montevecchi. 2000. Elevated numbers of flying insects and insectivorous birds in riparian buffer strips. Canadian Journal of Zoology 78: 740-747. Whitaker, D . M . and W.A. Montevecchi. 1997. Breeding bird assemblages associated with riparian, interior forest, and nonriparian edge habitats in a balsam fir ecosystem. Canadian Journal of Forest Research 27: 1159 - 1167. Whitaker, D . M . and W.A. Montevecchi. 1999. Breeding bird assemblages inhabiting riparian buffer strips in Newfoundland, Canada. Journal of Wildlife Management 63: 167- 179. Whitcomb, R.F., Robbins, C.S., Lynch, J.F., Whitcomb, B.L., Klimkiewicz, K., and D. Bystrak. 1981. Effects of forest fragmentation on avifauna of the eastern deciduous forest. Pp. 125 - 205 in Forest island dynamics in man-dominated landscapes. Burgess, R.L., and D . M . Sharpe (eds.). Springer-Verlag, New York Wiebe, K . L . and K. Martin. 1997. Seasonal use by birds of stream-side riparian habitat in coniferous forest of northcentral British Columbia. Ecography 21: 124 — 133. Wiens, J. A. 1989. The Ecology of Bird Communities. Volume 1: Foundations and Patterns. Cambridge University Press. Cambridge, UK. Wiens, J.A. 1994. Habitat fragmentation: island v landscape perspectives on bird conservation. Ibis 137: S97 - SI04. Worton, B.J. 1987. A review of models of home range for animal movement. Ecological Modelling 38: 277 - 298. Wright, A . L . , Hayward, G.D., Matsuoka, S.M. and P.H. Hayward. 1998. Townsend's warbler (Dendroica townsendi). In The Birds of North America, No. 333. A. Poole and F.Gill (eds.). The Birds of North America, Inc., Philadelphia, PA.  98  Zanette, L. 2000. Fragment size and the demography of an area-sensitive songbird. Journal of Animal Ecology 69: 458 - 470. Zanette, L . and B. Jenkins. 2000. Nesting success and nest predators in forest fragments: A study using real and artificial nests. The Auk 117: 445 - 454. Zar, J.H. 1996. Biostatistical analysis. Third Edition. Prentice Hall, New Jersey, USA.  99  03 O  S-l  >> T3  3  oo D  e ©  t  « cu  S (A CJ  v-T  c*J  fa  o o o  fa  Xi  o cd c_> to  00  13  00  3  « t/3  p<  D  D  Pi Pi  D  P<^)  Pi Pi  ID  D O fa Z Z fa fa fa fa  H  D  Pi p i  pi O  fa H fa H H  a cd  3cd  +->  B _© 08  a o  -*-> cd  > >-<  >• CU  S-  Xi Xi  <  o  9 KK  > fa 2 u O Pifaffi  .  5 j K  i  o  -<  2hjSSSzzoofa  <<B5fflfflUUUDQwOO  •s oo CD a,  00  O  o oo  CL>  -a  1  cu =  fa« T3 o S o  z, 00  CO  —  • r-H O  c ©  s s  cd i-  o U  ed =  '5  u  co  s  CJ  CO  ccS X )  o o  ccj  ccj  I CO  43  C8  ccj  o >.  w  £ Is 0 _g X  •a  -a  E l t; O  CO  CJ  VH  GO 00 S 4>  X)  c o o  O  1 o  4>  2 S  O fa  C3  s  o  K  cu  00  a -a  CN  X  cu  Q  S3  § fan  E  _  •~j  3 .bo -  C3  CJ  a 2 „ s p 2 cj "Ji  C  CD  C-3  i s  ^  C3  C3 ^3 w ^  "a •K CD  to  "C3 K O  cj  ca a  CD  a 2  CD  C3 C3  cj to  CJ  to S  03  s K o, a 2 o G 'G •S "8 "3 £ c5 s .<2 S, 3 a t ^ bq co be K tt5 cS:( J o 0 £ a, a, t<u L o a O oq cJ^U CD  CD  CD  CU  u  B  CD  CD  CD  a o cs "S •§ "Q 2 cj a o K s S a C3 5 s bo o, a S o c J bo - K -s -O cl-2 -i:  a  CD  CJ  ^ I, CJ  CD  CD  K  CD  101  O  D  D Pi  2 h Z Z  OH  . •  D . H V PH  U  5  .  U  O  o  D U , r rD D O U . H ' - h h r  •c  M  >> -o 3 o J3  102  r-  OS ON  os  ON ON  ON ON —  _ : T3 CO C  CO  o o o CN  ^co  *—<  ;  t co  <u 3 CO 0)  US'  ON ON ON  ->  O -o c  CO 0>  av  C Q <II Q ll ll  co  NO ON ON CO ^ H  ^  E cj  CO  C/3  W H  2 X  o  H  —  CM  CJ VN C OS >N ON  IZ3 O  II  w  CO 00 CO  ll  II  II  SO  f-  BO  c  „ « b  i_ o  yu  o o PNI  OS ON  OX  .O  O  X  cj O  u  u  .3  • 3  •5 J= 11  ON  o —  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Country Views Downloads
United States 30 10
China 17 6
Canada 16 0
France 5 0
Germany 3 10
Japan 2 0
Sweden 1 0
Latvia 1 0
Romania 1 0
Italy 1 0
United Kingdom 1 2
Netherlands 1 0
City Views Downloads
Unknown 17 29
Shenzhen 15 6
Ashburn 10 3
Wilmington 6 0
Cupertino 6 0
Fort Worth 3 0
Kelowna 3 0
Comox 2 0
Beijing 2 0
Denver 2 0
Thunder Bay 2 0
Tokyo 2 0
Mountain View 2 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090468/manifest

Comment

Related Items